vx
EPA
UnHadSlatM
Environmental Protection
Agency
Region 1
JFK Federal Building
Bo«on, MA 02203
Connecticut
Main*
MasaachuMtts
N«w Himp«hlr*
Vermont
Rhodi l«l»id
Planning & Mwagwiwni D(vi»on Planning, ArMlycis, & Qrantt Branch Daownbar 1988
Unfinished Business
in New England:
A Comparative Assessment
of Environmental Problems
Public Health Risk Work Group Report
-------
Unfinished Business in New England:
A Comparative Assessment of
Environmental Problems
Public Health Risk Work Group Report
United States Environmental Protection Agency
Region I, Boston, Massachusetts
December 1988
-------
Acknowledgments
This report and all the work that supports it would not have been possible without the
commitment and unselfish efforts of the Public Health Risk Work Group and its support staff.
Those individuals and their organizations are listed below.
Norman Beddows, Planning & Management Division
Tom D'Avanzo, Air Management Division
Sally Edwards, Planning & Management Division
Kim Franz, Water Management Division
Susan Green, Waste Management Division
Corrine Kupstas, Water Management Division
•Sarah Levinson, Waste Management Division
Mark Mahoney, Environmental Services Division
Stephen Perkins, Air Management Division, Chair
Jon Pollack, Air Management Division
Pi- Yun Tsai, Water Management Division
Catherine Tunis, EPA Headquarters
Margo Levine, Temple, Barker & Sloane
Julia Shepard, Temple, Barker & Sloane
Special thanks go to Phyllis Gould and JoAnne Hanna, Air Management Division, for
preparation of this report.
-------
Table of Contents
L INTRODUCTION 1-1
H. ANALYTICAL METHODOLOGY H-l
A. General Approach n-1
1. Ground Rules . n-1
2. Health Risk Criteria U-l
3. Overall Methodology D-2
4. Other Considerations n-3
B. Cancer Methodology n-4
1. Identify Representative Chemicals n-4
2. Characterize Potency n-4
3. Assess Exposure n-4
4. Characterize Risk H-5
5. Estimate Percent Covered and Uncertainty H-5
6. Results of analysis II-5
C. Non-cancer Methodology n-7
1. Identify Representative Chemicals n-7
2. Determine Severity H-7
3. Characterize Potency H-9
4. Assess Exposure H-9
5. Characterize Risk H-10
6. Estimate Percent Covered and Uncertainty H-11
7. Results of Analysis H-11
m. RELATIVE RANKING m-1
A. Cancer Risk m-1
B. Non-cancer Risk DI-3
C. Overall Risk UI-5
IV. OBSERVATIONS IV-1
-------
Table of Contents (continued)
APPENDIX: Problem Area Papers
Area#l.
Area #2.
Area #3.
Area #4.
Area #5.
Area #6.
Area #s 7,8,9.
Area #10.
Area #12.
Area #13.
Area #14.
Area #15.
Area #16.
Area #17.
Area #18.
Area #19.
Area#s20&21.
Area #22.
Area #23.
Criteria Air Pollutants
Acid Deposition and Visibility
Hazardous/Toxic Air Pollutants
Radon
Indoor Air Pollutants Other than Radon
Radiation from Sources Other than Radon
Discharges to Surface Waters
Discharges to Estuaries, Coastal
Waters, and Oceans from All Sources
Drinking Water
RCRA Waste Sites
Superfund Waste Sites
Municipal Waste Sites
Industrial Waste Sites
Accidental Releases
Releases from Storage Tanks
Other Ground-Water Contamination
Pesticide Residues on Food
& Pesticide Application
Lead
Asbestos
1
6
10
16
26
31
34
37
40
48
51
55
58
59
61
64
78
80
86
-------
I. Introduction
The Public Health Risk Work Group was charged with evaluating the residual cancer and
non-cancer risks posed by the 24 environmental problem areas being addressed in the Risk
Reduction Project (RRP) and then providing a relative ranking of residual public health risk. .
The work group's initial goal was to perform a quantitative analysis for each problem area.
However, since available data were often limited and not easy to compare directly, we relied
on a significant amount of professional judgment during the ranking process. The
conclusions reached by the work group therefore represent judgments and not verifiable fact.
Nevertheless, the work group feels confident that the rankings we produced represent relative
differences in residual public health risk across environmental problem areas.
The work group was composed of technical staff from each of the Region I program
offices and was supported by a representative from New England interstate environmental
agencies, a representative from the EPA Headquarter's Office of Policy, Planning and
Evaluation, and a contractor. Representatives were selected based on public health expertise
and specific program knowledge. The work group first discussed the methodology to be used
in assessing residual public health risk. Group members were then designated as problem area
leads according to their individual areas of expertise. Work group members, or contractors
under their direction, then prepared "plan of analysis" papers outlining how each problem area
would be analyzed to determine the residual risk it posed. The plans were then discussed by
the work group. With contractor assistance, the work group members then collected and
analyzed relevant data and prepared their analyses for consideration by the rest of the work
group. The work group met twice each month over a period of four months to develop its
methods and outline the problem areas, and then weekly for one month to discuss the
evaluation of each problem area and perform the ranking.
This report first discusses the methodology developed by the Public Health Risk Work
Group to evaluate the problem areas. We present our general approach and then discuss the
specifics of the cancer and non-cancer approaches. The next section presents the ranking
process for cancer and non-cancer effects, followed by the process used to generate the overall
public health risk ranking. A final section provides our observations on the difficulties we
encountered and other key lessons we learned in the process. The appendix to this report
presents in greater detail the methods, data, and conclusions tor each of the problem areas.
-------
Analytical Methodology
Since there is no single, EPA-endorsed, off-the-shelf method for conducting a
comparative public health risk evaluation, the work group set out to define a process to govern
our analysis and deliberations. This section describes the work group's ground rules, the
general method we developed, and more specific details on the cancer and non-cancer
methodologies.
A. General Approach
Ground Rules
At the outset of the project certain ground rules were established to provide consistency
across the three work groups participating in the RRP. Two of these ground rules were very
important to the way the Public Health Risk Work Group approached the analysis. First, we
considered each problem area assuming current levels of control and compliance. This means
we considered the residual risk posed by the problem area and not the risk that might be posed
if controls were changed or if compliance rates dropped.
This led to a discussion of what constitutes controls. The work group expanded the
obvious list of pollution control equipment and pre-treatment to include fencing of Superfund
sites or other waste facilities. We recognize that trespass can and will occur and have
evaluated that risk, but we have assumed that one cannot currently drill a drinking water well
on-site at such a facility. We did not include announced "advisories" restricting fishing or
swimming as controls since they depend on the awareness and cooperation of the public for
effectiveness.
The second important ground rule was that we consider current emissions and their effects
now and in the future. We sought the most recent data available to define those current
emissions with the understanding that data as much as three to five years old might be the best
available. In cases where effects happen over many years, we did not discount the future
effects.
Health Risk Criteria
With the ground rules set, we next considered the types of public health risk to analyze.
We decided to first look at cancer and non-cancer risks separately, postponing until later a
discussion on how we would combine the two. Within these two categories we agreed to
examine both individual risk and population risk. Individual risk was defined as die risk
posed to an individual potentially exposed to the chemicals under consideration at what was
determined to be an average concentration. Population risk considered the risk posed to all
-------
people determined to be exposed to this concentration in Region L In both the individual and
population risk analyses, we agreed to consider realistic, average exposures, not worst-case
exposures. Thus we have not explicitly considered the maximum exposed individual Finally,
we agreed to consider both chronic and acute effects, focusing on the more typical exposure
for each problem area.
Overall Methodology
We then developed an overall methodology for assessing the types of public health risks
described above. In developing the methodology, we used the approach taken in the National
Comparative Risk Project (NCRP) as the point of departure and the considered alternatives at
each step. While our method ended up looking quite similar to the NCRP method, we did not
take it as a given at the outset The broad outline of our methodology is described below,
with further details on the cancer and non-cancer approaches presented in succeeding sections.
The first step was to identify a short list of chemicals and the associated possible exposure
pathways to represent each problem area. We relied on individual expertise to identify the
chemicals, sources, and pathways. The chemicals were selected on the basis of how well they
reasonably characterized the problem and the extent to which data were available. The
number of chemicals studied for any problem area ranged from 1 (e.g., lead, asbestos, radon)
to 10.
The second step was to characterize the potency or dose-response relationship for each
chemical. While this characterization was straightforward in the cancer analysis, developing
an analogous approach for non-carcinogens was complicated, and the reader is referred to the
non-cancer methods section for further details.
To characterize the severity of the effects, we added an additional step to the non-cancer
methodology. This was necessary because, unlike cancer, there are many different types of
non-cancer effects, and they vary widely in severity from one chemical to the next These
systemic toxic effects include cardiovascular, developmental, immunological, kidney, liver,
gastrointestinal, neurotoxic/behavioral, reproductive, respiratory, and other effects. The work
group embraced die NCRP approach of using a severity index to sort out the type of
impairment and the part of the body affected. Further details are provided in die non-cancer
methods section below.
The third step was to assess exposures. This involved determining the likely
concentration levels at which exposure occurs, the resulting doses, and the likely population
exposed. We focused cm average rather than worst-case exposures. The exposure assessment
was based on available ambient monitoring data, modeled estimates, and available data on
actual incidence. The method for any problem area depended on the availability of actual data
or an appropriate modeling approach. Rather than selecting one method, the work group
sought the best available exposure estimate within each problem area.
H-2
-------
The next step was to combine these factors to characterize the individual and population
risk for each chemical within each problem area. In the cancer analysis we quantified
individual risk a"d a simplified estimate of annual cancer incidence. In the non-cancer
analysis we developed a combined score based on the exposure, potency, and severity scores.
In problem areas with more than one chemical, the risks were combined to arrive at an overall
cancer risk, while for non-cancer risk the highest score among all chemicals was used.
The final step in the process was to develop estimates of the percentage of the problem
area covered and the uncertainty of our risk estimate. These were described qualitatively as
high, m^tinm, or low. The estimate on percentage of problem covered was intended to help
us account for the portion of the issue not analyzed when it came time to rank the problem
areas. We decided not to use the estimate of the problem area covered to mathematically scale
the risk estimate. Rather, we used the estimate to provide a qualitative basis for adjusting the
risk estimate relative to other problem areas. The uncertainty estimate was used to
characterize our confidence in the data. We used best-guess estimates of uncertainty to help
provide some bounding of our risk estimate..
Other Considerations
The work group decided to refine the list of environmental problem areas. First, we
eliminated two of the problem areas from further analysis. We eliminated Wetlands/Habitat
Loss (#11), since there clearly were no direct public health effects associated with this
problem area. We also eliminated Lakes, Ponds, and Impoundments (#24), since we could
not readily distinguish any public health risks not already covered by Discharges to Surface
Waters or Estuaries (problem areas #7 through #10).
Discharges to Surface Waters from Industrial Point Sources (problem area #7), from
Publicly Owned Treatment Works (POTWs) (problem area #8), and from Nonpoint Sources
(problem area #9) were analyzed as a group due to our inability at the beginning of the
analysis to separate out the contributions of each problem area to the water quality data
collected. Once the risks were characterized and the chemicals posing the greatest risks were
identified, an attempt was "»
-------
B. Cancer Methodology
This section describes die cancer methodology in a step-by-step fashion. Steps commc
to the cancer and non-cancer approaches were described above and will only briefly be
reviewed, while details unique to the cancer methodology will be given a more thorough
discussion.
Stepl: Identify Representative Chemicals
We selected from 1 to 10 cancer-causing chemicals for each problem area and identified
the likely sources of these chemicals in Region L The chemicals were selected based on our
estimate of their contribution to the problem and the availability of data on exposures and
health effects. At this point in the cancer analysis, three problem areas were categorized as
not posing cancer risks. Criteria Ah* Pollutants (#1), Acid Deposition and Visibility (#2), and
Lead (#22) were so characterized because of the existing presumption that the substances of
concern are not currently considered to be carcinogens.
Step 2: Characterize Potency
The purpose of this step was to determine for each of the chemicals the relationship
between the dose and the probability of developing cancer. This relationship was used to
predict the level of risk associated with a given exposure. While there are various models for
assessing carcinogenic potency, EPA has only approved the linearized multistage model.
Using thiy model, it is assumed that at low doses the dose response curve is linear flnd that the
slope goes through the origin (Le., there is no threshold effects level for carcinogens).
Therefore the simplified equation "Incremental Cancer Risk = Potency times Dose" can be
used to determine an upper-bound probability of incremental cancer risk (Le., cancers that
would be expected above the background rate) for low doses of carcinogens. The woric group
adopted cancer potency factors derived from this model, where available. The potency factors
were generated by EPA's Carcinogen Assessment Group (GAG).
Step 3: Assess Exposure
The purpose of this step was to estimate the concentration levels at which exposure
occurs, identify the pathways and estimate doses, and estimate the size of the exposed
population. The assessment was based on average, not worst-case, exposures and made use of
available ambient monitoring or incidence data or modeling results. In most cases, exposures
were assumed to be constant over a 70-year lifetime. At this point in the analysis, Accidental
Releases (#17) were assumed to pose low cancer risks because chronic exposures were not a
reasonable assumption. It is possible that a severe short-term exposure could cause cancer
later in life, but we assumed that acute non-carcinogenic effects were of most concern for this
problem area.
H-4
-------
Step 4: Characterize Risk
In this step, exposure dose and potency estimates were multiplied together to calculate
upper-bound individual excess cancer risks for each chemical. The individual risk was then
multiplied by an estimate of the exposed population to give an estimate of the number of
cancer cases expected over a period of 70 years. This number was then divided by 70 to
provide an estimate of annual cancer incidence. The group agreed that it was reasonable to do
this calculation to provide a relative estimate of annual cancer risk. For problem areas with
multiple chemicals, die individual and population risks were each summed over die chemicals
if it was assumed that individuals were exposed to the mixture.
Step 5: Estimate Percentage Covered and Uncertainty
Consistent with the overall methodology, qualitative estimates (Le., high, medium, low) of
the percentage of the problem covered (i.e., to what extent were we able to extrapolate to the
Region as a whole from our analysis of a portion of die problem) and the uncertainty
associated with the data were made for each problem area.
Step 6: Results of Analysis
A summary of the results of the cancer analysis is presented in Table n-1. The table
displays the estimated individual risk, population risk (in cancers per year), the percentage of
problem covered, and die estimated uncertainty.
For details on the data supporting the summary table, the reader is referred to the specific
papers on the problem areas contained in die appendix. For some problem areas, no data
appear because no quantitative analysis was performed. These problem areas were handled
qualitatively by group consensus when it came rime to perform the ranking. Problem areas
#7, #8, and #9 are shown as a total risk from discharges to surface water. A qualitative
attempt to differentiate between the three problem areas was deferred until the ranking.
H-5
-------
Table 11-1
Summary of Cancer Analysis
Problem Ana
1. Criteria Air Pollutants
2, Add Deposition and Vislbity
3. Hazardous/Tone Air Pollutants
4. Radon
5. Indoor Air Poflutanta Other than Radon
Individual
Risk
1E-4
2E-4to4E-3
Oto1E-2
Pa
Population of
Risk* C
NO CANCER RISK
NO CANCER RISK
-10
375-1500
1-60
raantago
Problem
iovarad UnewUlnty
L
H
H
H
L
M
6. Radiation from Sources Other than Radon (non-ionizing) NO QUANTITATIVE ANALYSIS PERFORMED
r7. Industrial Point Source Discharges to Surface Waters
! 8. POTW Discharges to Surface Waters
1 9. Nonpolnt Source Discharges to Surface Waters
10. Discharges to To Estuaries, Coastal Waters,
and Oceans from AO Sources
1£ Drinking Water
13. RCRA Waste Sites
14. Superfund Waste Sites
15. Municipal Waste Sites
16. Industrial Waste Sites
18. Releases from Storage Tanks
IB. urar ufDuno-wanr ccnttnunaoon
21. Pesticide AppOcaflon
22. Lead
23. Asbestos
1E-2
1E-4
1E-4to1E-5
1E-6
1E-5
1E-6
1E-5
1642
90
20
1-10
1
FEW
FEW
i nui RAMROD max
L
M
M
H
H
M
M
H
H
H
H
H
H
H
NO QUANTITATIVE ANALYSIS PERFORMED
1E-7
2E-4
••b^V
1E-4to1E-7
-1
917
W 1 V
-1
NO CANCER RISK
185
L
i
!•
L
L
H
H
H
•Population risk expressed as excess cancers per year.
Note: L-kwr. M-medhm. and H-hkjh.
n-6
-------
C. Non-Cancer Methodology
This section describes die non-cancer methodology in a step-by-step fashion. Steps
common to the cancer and non-cancer approaches were described under the general approach
and will only briefly be reviewed while details unique to the non-cancer methodology will be
given a more thorough discussion.
Step 1: Identify Representative Chemicals
We selected from 1 to 10 non-cancer-causing chemicals for each problem area and
identified the likely sources of these chemicals in Region L Unlike the cancer analysis, each
of the problem areas was characterized as posing some non-cancer risk.
Step 2: Determine Severity
The purpose of this step was to characterize the type and different severities of health
endpoints of the selected chemicals. The work group considered a number of alternative
schemes before adopting the severity index approach used in the NCRP. The NCRP approach
will be outlined below, and the interested reader is referred to Appendix n of EPA's
Unfinished Business for complete details.
In the NCRP approach, health effects were classified into 1 of 11 categories that were then
broken down into 106 different health endpoints. A severity index basically ranked health
effects by how threatening they were to the viability of the organism. This led to the
development of a 7-point severity scale. The 106 separate endpoints were given scores from
1 to? based on the index. The 7-point index was later collapsed to a4-point scoring scheme
by combining severity scores 1 and 2 into one score, severity scores 3 and 4 into another
score, and severity scores 5 and 6 into a thud score. The 4-point scale better fit the NCRP's
sense of the precision of the data with which they worked. Table n-2 shows the distribution
of the endpoints on the 4-point scoring scale. It should be noted mat this severity scoring
approach has not been endorsed by EPA and is quite controversial. However, the NCRP
recognized the need to make such rough judgments in order to analyze non-cancer risk. They
did feel that mere is a large difference in severity represented by a one point difference in the
severity score.
The work group spent a significant amount of time discussing the pros and cons of mis
approach before adopting it We decided that it best served the purposes of the work group
and agreed that it would have taken a great deal of time and effort to improve the NCRP
method. We also lacked sufficient expertise to devise a better alternative.
H-7
-------
Table 11-2
Distribution of Non-Cancer Health Endpolnts
on the Severity Scoring Scale
Seen ml Scon -2
-hopes (non-infectious)
-hvcr-mcreBsed euyinea -dmhiHtMy-nwnomMy
-reduced caracal senriliviiy -aggravation of asthma
•sensory irritation -increased iwpiruoiy disease
-giant cell formation -broachoconsirictions
-nasal itritation -increased respiratory infections
-nasal crlhilflr j*»fa"fafl -gastrointestinal
-eye hritation
^Bcotu mottling
Scon * 2 -aggravation of angina
•bone manow hypoplasia
-kidney-bjpeiplasia ^diiey-nibular degeneration
-kidaey-lqfpenraphy -jiepatttu A
-liver-hweased weight -AChE mhibition
Recreated sensory perception -increa»ed spontaneous abortion
-decreased testicular weight -Pontiac fever
-uterine hypoplasia -alveolar collapse
-ptdmonoy irritation -fibrosis
•«asalulcenuion -hmgstmc
-mucosal atrophy -cataracts
-decreased mid^qriratory flow rates -fetonaicity
Cental erosion -kidney-!
4ejshmamasis -liver-necrosis
-symptomatic effectt (headache) -iMnung disabilities
-increased blood pressure -neuropathy
-low both weight -convulsions
-decreased hemc production -aspenma
-impaired heme synthesis -emphysema
-increased infections -puhnonary edema
ey atrophy
-jaundice Scon m 4
-multagenicity-cytogaiic
-retinal disorders -increased heart attacks
-tremors -teratogenicity
.p^t hTT1""*h>t'"n lmm« -muiagenicity-heredilay disorders
-retafdation
ato danuute — iDOrtauty
resorptions
"OEOOCutuS
•puhnonaryimpninneTit
.bag injury
Note: l=less severe; 4emosl severe.
n-8
-------
Step 3: Characterize Potency
The purpose of this step was to determine the relationship between the dose of a substance
and the likelihood that the dose will produce an adverse health effect Most non-carcinogenic
dose-response functions are thought to involve a threshold below which it is unlikely that an
adverse health effect will occur because, up to a certain dose, the human body is able to
detoxify the chemical and repair any damage. The work group, after considerable discussion,
decided again to adopt the NCRP approach. (The interested reader is referred to Appendix n
of Unfinished Business for complete details.) The NCRP developed an "individual exposure
ratio" that was calculated as the actual exposure dose divided by the Reference Dose or RfD.
(The RfD is based on a no or lowest observed effects level in the study, divided by a safety
factor normally ranging from 10 to 1,000.) The higher the exposure dose above the RfD, the
higher the individual exposure ratio. At exposure doses below the RfD, we expect little
likelihood of adverse effects occurring, while at levels above the RfD it is reasonable to
assume that an adverse effect may occur in at least some sensitive population. This individual
exposure ratio therefore provides some indication of the potency of these non-carcinogens
relative to the estimated exposure dose. A four-point ratio score that reflects order of
changes in the individual exposure ratio was then developed. This broad four-
point scoring system once again reflected the NCRP's sense of the precision of the data
available to them. The ratio scoring scheme is presented in Table n-3 (Section A).
For more than half of the problem areas, the exposure was judged to be below the RfD,
producing an individual exposure ratio less than 1. In these cases no ratio score was assigned
and it was assumed that there was negligible risk. This allowed us to keep separate those
problems where there was a ratio in the 1 to 10 range from those where the ratio fell below 1.
For some problem areas, particularly in air programs, there are no established RfDs. In
these instances the work group member investigating the problem developed a surrogate RfD
after investigating other forms of existing standards, the values considered by the NCRP, and
other suggestions from experts at EPA Headquarters.
Step 4: Assess Exposure
The purpose of this step was to estimate the likely concentration levels at which exposure
occurs, the resulting doses, and the likely populations exposed. The assessment was based on
average exposures and made use of the available data, whether measured or modeled. After
much discussion, we adopted a modification of the NCRP population scoring approach for
assessment.
The NCRP approach convened the estimated population exposed to a chemical into a
from 1 to 4. One point of change in the score represents a difference in exposed
population of two orders of magnitude. Thus errors of up to a factor of 10 in estimating
exposed populations would make little difference in the final result The work group
modified the population ranges associated with each score to reflect populations more
appropriate for the region. However, we maintained the NCRP's two orders of magnitude
change in population between each score. Table n-3 (Section B) presents the work group's
population scoring scheme.
n-9
-------
Table 11-3
Scoring System for Non-Cancer Risks
A. Ratio Score dorivod as foQows:
lual Exposure Rate
1 1to10
2 10 to 100
3 100101,000
4 >1,000
B. Population Score derived as follows:
Score Number of People Exposed
I I to 100
2 100 to 10,000
3 10,000 to 1,000,000
4 >1,000,000
C. Ratio Score derived from incidence data as follows:
Score fCases/OPeople Exposed Individual Exposure Ratio
1 <1.00E-06 1to10
2 1.00E-OS-1.00E-04 10 to 100
3 1.00E-04-1.00E-02 100 to 1000
4 >1.00E-02 >1000
In cases where incidence data were available, a different approach was used The
incidence data were converted to a population score and a ratio score to be consistent with the
remainder of the non-cancer approach. A ratio score was calculated as the annual incidence
divided by the estimated population at risk. Table H-3 (Section C) shows the ranges
associated with each score and the individual exposure ratio for comparison. The population
at risk was then converted into a population score.
Step 5: Characterize Risk
In mis step the severity, potency, and exposure evaluations were combined to estimate
non-cancer risk for each chemical and problem area. Since the work group had relied on a
scoring system for the three components, we could not estimate individual and population risk
in the same manner as we had in the cancer analysis. Instead, we had three separate scores for
each chemical of concern in each problem area. We could combine the three scores in various
ways to give us a relative sense of the differing risks posed by the problem areas.
n-io
-------
In characterizing the population risk, the work group considered various ways of
combining the severity, individual exposure ratio, and population scores. We finally decided
to sum the three scores for each chemical analyzed in a problem area and assign the highest
score within a problem area to that problem area. Other combination schemes, including
averaging across chemicals within a problem area, rnarfe. little or no difference in the relative
scoring. A surrogate score for individual risk was developed by adding the severity and
individual exposure ratio scores for each chemical and selecting the maximum sum within a
problem area.
As mentioned in the potency discussion above, the estimated average exposure was below
the RfD for about half the problem areas. Therefore, we did not assign an individual exposure
ratio score. In characterizing the risk for these problem areas, we did not form a combined
score but rather grouped these problem areas in a low risk category.
Step 6: Estimate Percentage Covered and Uncertainty
Consistent with the overall methodology, qualitative estimates of the percentage of the
problem covered (ie., to what extent were we able to extrapolate to the Region as a whole
from our analysis of a portion of the problem) and the uncertainty associated with the data
were made for each problem area.
Step 7: Results of Analysis
A summary of the results of die non-cancer analysis is presented in Table n-4. For each
problem area the table displays the severity score, the individual exposure ratio or potency
score, the population score, the percentage of problem covered, and the estimated uncertainty.
For the problem areas where the individual exposure ratio was less than one, the individual
and population risks are categorized as low rather than by a score. For details on the data
supporting each of these scores, the reader is referred to the specific papers in the appendix on
each problem area.
For some of die problem areas, no data appear because no quantitative analysis was
performed. These problem areas were handled qualitatively by group consensus when the
ranking was performed. As in the cancer analysis, problem areas #7, #8, and #9 are shown as
a total risk from discharges to surface water. A qualitative attempt to differentiate between the
three problem areas was postponed until the ranking was performed.
n-n
-------
Table 11-4
Summary of Non-Cancer Analysis
PIOM6IH AM8
Severity
Sooro
Scorn
Percentage
Population of Problem
Score Covered Uncertainty
1. Criteria Air Pollutants
2. Acid Deposition and Visibility
3. Hazardous/Toxic Air PoOutants
4. Radon
& Indoor Air Poflutants Other than Radon
a Radation from Sources Other man Radon
(norrionUng)
7. Industrial Point Source Discharges
to Surface Waters
a POTW Discharges to Surface Waters
9. Nonpoint Source Discharges
to Surface Waters
10. Discharges to Estuaries. Coastal Waters.
and Oceans from AH Sources
M Drinking Water
13. RCRA Waste Sites
14. Superfund Waste Sites
15. Municipal Waste Sites
16. Industrial Waste Sites
17. Accidental Releases
ia Releases bom Storage Tanks
19. Other Ground-Water Contamination
20. Pesticide Residues on Food
21. Pesticide Application
22. Lead
23. Asbestos
234
334
NO QUANTITATIVE ANALYSIS PERFORMED
LOW NON-CANCER RISK
3 1 3
LOW NON-CANCER RISK
M
LOW NON-CANCER RISK
2 3
LOW NON-CANCER RISK
LOW NON-CANCER RISK
LOW NON-CANCER RISK
LOW NON-CANCER RISK
3 2
3 NOTSCORED
3 2
LOW NON-CANCER RISK
2 2
3 4
LOW NON-CANCER RISK
2
3
4
3
3
M
M
L
L
L
H
L
H
M
M
H
M
H
H
H
L
Note: LB|OW, Mafliedium, and Hoftigh.
n-i2
-------
in. Relative Ranking
Once the risk analyses were completed, the work group met to perform a relative ranking
of the problem areas for cancer and non-cancer risk. We considered options ranging from
ordinal ranking to grouping into a few categories. We were not comfortable with developing
an ordinal ranking because we did not believe our analysis provided us with anything
approaching the level of precision needed to perform an ordinal ranking. We agreed to set a
target of ranking the problems into five categories.
The ranking process was primarily influenced by the outcome of the two risk
characterization efforts. It was supplemented by significant discussion and group consensus.
For the problem areas where data gaps had prevented the calculation of an individual or
population risk estimate, we worked with whatever data we had and our sense of the other
problem areas to reach a consensus relative ranking.
Once the cancer and non-cancer rankings were completed, we embarked on the most
difficult venture of all-developing an overall public health risk ranking. This chapter
provides details on the process and the results of each of the three ranking exercises.
A. Cancer Risk
The work group's first cut at a relative ranking of cancer risk was based on the population
risk which we had expressed as an annual cancer incidence. We had already created our
lowest ranking group when we decided that Criteria Air Pollutants, Acid Deposition and
Visibility, Accidental Releases, and Lead posed no or low cancer risks. Looking at the range
of annual cancer incidence predicted in the analysis, the work group developed four additional
ranges to use in the preliminary ranking. The worst (highest) ranking, Category 5, was defined
as having an annual cancer incidence greater than 250 in the region. Category .4 was defined
by the range of 100 to 250 annual cancers, Category 3 by 10 to 100 cancers, and Category 2
by 1 to 10 cancers. Category 1 contained the problem areas that posed no cancer risk.
The next step, for which we had very little data and high uncertainty, involved ranking
Radiation from Sources Other than Radon and Releases from Storage Tanks. We elected to
rank both in Category 2 based on a qualitative sense that there was some small risk. We also
attempted to separate the risk from discharges to surface waters into the three separate
problem areas. The risk estimates were driven by ingestion of fish contaminated with PCBs.
The work group thought that Industrial Point Sources and Nonpoint Sources were the major
contributors to this problem. POTWs were judged less likely to contribute to the risk, so we
reduced the ranking for POTWs by one category. This determination was based on the types
of contaminants contributing to the major risk.
-------
At this point we reviewed the rankings to see if they made relative sense. Our review
considered the individual risk, the percentage of problem covered, and the uncertainty. In mis
process, we revised the ranking of three problem areas. Hazardous/Toxic Air Pollutants,
which was between Categories 2 and 3, was placed in Category 3 based on high individual
risk and the low percentage of problem covered. Pesticide Residues on Food was dropped to
Category 4 because of the work group's belief that the analysis scheme, based on NCRP data,
had overstated the extent of die problem. Similarly, we thought the Asbestos analysis had
overstated the problem and dropped it into Category 3.
The final results of the cancer ranking are presented in Table ffl-1. Within each category,
problem areas are listed in the order they were originally listed for the project; no relative
ranking within a category was performed or should be inferred.
Table 111-1
Cancer Relative Risk Ranking
(unranked within categories)
High Category S
Risk Radon
Category 4
Category 3
Hazardous/Toxic Air PoOutantt
Indoor Air Pollutants Other than Radon
Industrial Point Source Discharges to Surface Waters
Nonpolnt Source Discharges to Surface Waters
Discharges to Estuaries. Coastal Waters, and Oceans from All Sources
Drinking Water
Category 2
Radiation (torn Sources Ottwr than Radon (norvtonizino)
PO7W Dischargee to Surface Waters
RCRA Waste Sites
Superfund Waste Sites
Municipal Waste Sites
Industrial Waste Sites
Releases from Storage Tanks
Other Ground-Water Contamination
•• J—«-J—. AflM^^Mtllj^M
resociOB Appucaoon
Category 1
Criteria Air Pollutants
Add Deposition and VWbifiry
Risk Lead
m-2
-------
B. Non-Cancer Risk
The work group's first attempt at a relative ranking of non-cancer risk was based on the
population risk estimate which we had expressed as a total score. The work group used the
scoring range to develop five categories. Problem areas scoring 10 or more were put in
Category 5, areas scoring 8 or 9 went in Category 4, areas scoring 6 or 7 went in Category 3,
areas scoring 1 through 5 went in Category 2, and the problem areas that had been classified
as "low" went in Category 1. The next step involved ranking Releases from Storage Tanks,
for which we had little risk information. We elected to rank it in Category 3 based on a
qualitative sense that the risks were on the same scale as other problem areas in this category.
We also attempted to separate out the risks from discharges to surface waters. The risk
estimates were driven by ingestion of fish contaminated with mercury and lead. The work
group thought that Industrial Point Sources would likely be the major contributor, followed
by Nonpoim Sources and then POTWs. POTWs and Nonpoint Sources were dropped down
one category at this point.
The nature of the scoring scheme resulted in no problem areas listed for Category 2 and 12
problem areas listed in Category 1. The work group spent time seeing if it could subdivide
this low-risk category into two groupings, but ultimately we could not The non-cancer
ranking thus shrunk to a four-category scale.
The work group then reviewed each of the problem areas, considering the individual risk,
percentage of problem covered, uncertainty factors, and other information that might have
come to light after the original analysis was conducted. This process resulted in revised
scores for Criteria Air Pollutants (#1) and Acid Deposition and Visibility (#2). Criteria Air
Pollutants had originally been ranked in Category 4, with a score of 9. This score was driven
by increased asthmatic attacks and reduced activity in susceptible populations as a result of
widespread violation of the Ozone National Ambient Air Quality Standard. Additional
information indicates that levels at or near the standard result in respiratory effects in healthy
adults and that repeated exposures heighten sensitivity. The summer of 1988 also showed an
increase in days exceeding the standard over a wider area than in previous years. Upon
consideration of this information, the work group moved Criteria Air Pollutants up to
Category 5.
Acid Rain and Visibility had originally been ranked in Category 5, with a score of 10.
This score was driven by increased hospital admissions due to respiratory illness resulting
from inhalation of acidic aerosols, primarily sulfates. However, significant uncertainty
surrounds the health effects data, which rely heavily on epidemiological studies. This
uncertainty led the work group to lower the relative ranking to Category 4. Health effects
from total paniculate exposures (including the acid aerosols) would also support a ranking in
Category 4.
The final results of the non-cancer ranking are presented in Table ffl-2. Within each
category, problem areas are listed in the order they were originally listed for the project; no
relative ranking within a category was performed or should be inferred.
m-3
-------
Table 111-2
Non-Cancer Relative Risk Ranking
(unranked within categories)
High Category 6
Risk Criteria Air Pollutants
CategonM
Acid Deposition and Visibility
Industrial Point Source Discharges to Surface Waters
DrinUnoWatsr
Other Ground-Water Contamination
3
Indoor Air PoOutants Other than Radon
Nonpohit Source Discharges to Surface Waters
Releases from Storage Tanks
Category 2/1
HuHidous/ToxIc Air PoSutsnts
•^—•—
nHUUII
Raolatton from Sources Omer than Radon (non-tordzing)
POTW Discharges n Surface Waters
Discharges to Estuaries, Coastal Waters, and Oceans from AD Sources
RCRA Waste Sites
Superfund Waste Sites
Municipal Waste Sites
Industrial Waste Sites
Low Pesticide floilduei on Food
Risk Asbestos
m-4
-------
C. Overall Risk
The final work group task was clearly the most difficult one: developing an overall
relative public health risk ranking. The work group spent more than two meetings on this task
and discussed many ways to approach the problem. We could not develop a clear-cut method
for combining the cancer and non-cancer risks. Instead, our ranking grew from a group
consensus-building process and ultimately represents our best collective professional
judgment
One path we examined involved using explicit approaches to combine the cancer and non-
cancer rankings. These approaches included averaging the rankings, using various weighted
averages, or using the highest of the cancer or non-cancer rankings to characterize a problem
area. All of these approaches are fraught with the difficulty of comparing very different
health effects. It is ultimately a personal value judgment whether a lung cancer case later in
life is better, equivalent, or worse compared with a lifetime with learning disabilities because
of childhood lead poisoning. Not surprisingly, the work group members' judgments in
evaluating these types of questions differed greatly.
The other major path we examined involved attempting to suspend judgment on the
thorny questions posed above, focusing instead on those problem areas that were likely to be
causing the worst public health problems in the region. This did not mean turning our backs
on the results of our earlier analysis. Instead, it meant considering to what extent the problem
area posed risks beyond levels that might be broadly portrayed as acceptable. This approach
relies on the fact that, as regulators, we have set standards for many substances, both
carcinogens and non-carcinogens, and have decided that exposures above these levels are
unacceptable. We could then consider the level and extent of these unacceptable exposures in
grading the problem areas. Using this approach, we would not be explicitly equating cancer
and non-cancer effects but rather would be saying that both can be bad above certain exposure
levels.
The work group ended up using a combined approach. Each member prepared his or her
own individual ranking, dividing the problem areas into four or five groups. Each person was
free to devise his or her own approach to combining the cancer and non-cancer rankings.
Many people used some explicit scheme as a starting point and then made adjustments based
on their professional judgments. Other rankings were based primarily on well-informed
judgments but were not explicitly driven by the cancer and non-cancer rankings. As one might
have expected, all of the individual rankings were different.
We then analyzed the individual work group member's rankings and identified the
problem areas that fell into the same ranking category in the majority of the rankings. About
two-thirds of the problem areas were thus sorted into five ranking categories. (The remaining
problem areas had split rankings, falling between the categories sorted out above.) We then
reviewed the relative ranking of the problem areas that had been placed in the same categories
in the majority of die individual rankings. We discussed each problem area, reviewing the
data we had analyzed and its uncertainty and comparing it with other problem areas in the
same ranking category. After these discussions, none of the initial problem area rankings was
changed. This process also provided a good benchmark for ranking the remaining areas.
m-5
-------
For the remaining problem areas, we used a similar process of reviewing the available
information for each problem area and comparing the risk to the risks of problem areas
already ranked. During the discussion we considered creating a sixth ranking category. We
rejected this idea because we did not believe the data available allowed us to make that many
gradations with confidence.
The most difficult pan of the process was agreeing on the highest risk category. We had
used Radon, which had received the highest cancer risk rating and also showed some small
non-cancer risk, as the benchmark for the highest risk category. We discussed at length
whether Criteria Air Pollutants and Lead, which had received the highest non-cancer risk
rating but low/no cancer risk, should join Radon in the highest overall risk category. We
ultimately did rank them in the same category. We recognize that the health effects posed by
each of these problem areas are very different and difficult to equate. However, we can
comfortably state that these three problem areas pose the most significant residual public
health risks in the region and that they stand apart from the other problem areas in this respect
Once we determined the high-risk category, the remaining problem areas fell fairly easily
into place. The final ranking reflects the work group's consensus. While some individuals
might have preferred to see a particular problem area shifted by one category, as a work group
we reached a point where no person was uncomfortable with the final result We agreed that
the uncertainty in the process made it difficult to draw very hard lines between adjacent
ranking categories, especially in the middle of the range. At the same time we are confident
that none of the problem areas were off by two ranking groups.
The final relative ranking of residual public health risks posed by the problem areas is
displayed in detail in Table ffl-3. The problem areas have been placed into five categories of
decreasing magnitude. Within each category, the problem areas are listed in the order in
which they were originally listed for this project No relative ranking within a category was
performed and none should be inferred. Within the table, the first column lists the problem
areas and the second column indicates the chemicals analyzed. The third column summarizes
our findings in both the cancer and non-cancer analyses and other relevant comments. A brief
summary, showing only the groupings of the problem area titles into the five categories, is
displayed in Table ffl-4.
Some of the major conclusions of the Public Health Risk Work Group include the
following:
• Indoor radon exposure, which is associated with lung cancer, poses a significant
residual health risk in Region L The estimated excess annual cancer cases
associated with radon exposure may be up to an order of magnitude greater than
those for other problem areas that pose a cancer risk.
• Lead exposure from all sources poses a significant public health threat in New
England, particularly to young children who may suffer from learning disabilities
as a result of mild to severe lead poisoning.
m-6
-------
• Exposure to ground-level ozone concentrations above the National Ambient Air
Quality Standards occurs frequently in the summer in Region I and is associated
with respiratory ailments that may affect large populations in New England.
• Pesticide residues on food are a significant pubUc health concern in Region L Hie
analysis available to us simply scaled a national estimate down to a regional
estimate, based on a straight ratio of regional to national population. The absence
of regional data made us reluctant to place it in the highest category. Nevertheless,
pesticide residues on food consumed in New England are a significant concern.
• Separating the residual public health risk posed by the three types of discharges
to surface waters was difficult After much discussion and input from Water
Management Division staff, we agreed to rank these problems as follows:
industrial point sources in Category 4, nonpoint sources in Category 3, and
POTWs in Category 2. This determination was based on the types of
contaminants contributing to the major risk.
• Most of the waste-related problems, including RCRA, Superfund, municipal, and
industrial waste sites, pose low residual public health risks. These problem areas
primarily show up as ground-water contamination problems that may or, more
likely, may not be affecting drinking water supplies today. While these sources
are contaminating aquifers that may be valuable future sources of drinking water,
current exposures involve relatively low concentrations and relatively small
populations.
• The work group agreed that, although our risk analyses would indicate that
municipal landfills pose risks similar to Superfund sites, there is great uncertainty
related to municipal landfills and the problem would greatly benefit from a more
in-depth analysis.
m-7
-------
Table 111-3
Relative Residual Public Health Risk Ranking
(unranked within categories)
ProbtanAfM
Comments
Criteria Air Poautanti
Add Deposition and
Visibly
Indoor Air Pollutants Oher
than Radon
Industrial Point Source
Discharges to Surface
Waters
Drinking Water
Ozone
^--«— •• 1-«-
MUDon Monoxne
PardculalB Matter
Radon
AddAerosds-Sulfates
and Nitrates
Carbon Monoxide
Nitrogen Oxides
Paniculate Matter
Formaldehyde
Benzene
Carbon Tetraohloride
Db^bfl»la«aMh C«AM^^
rnBIBHie BMMt
Cntofdane
Tobacco Smoke
Tnchtaraetnytone
TetradHoraethylene
PCB».Dkixin. Mercury
Total Trthatomethanes
OTHM). Arsenic,
RacSonudJdes, Nitrates,
Bacteria, Pathogens
Cancer Categoryl-Assumed no cancarrisk since all
criteria air pollutants ere currently considorod non*
carctn
Non-cancer Category 5-Driven by targe-scale exposure to
high ozone levels across the region
is cancers prodded
Cancer Category 6-Up to 1,600 excess ca
annually from exposure in homes
Non-cancer Category 2/1-Assumed low; some non-cancer
effects likely
Cancer- Category l-iAssumed no cancer risk since chemical
is not cunenfly oonslctefed caicrnogenfc
Non-cancer CategoiyS-OrivenbyingestionofsoDor
inhalation of lead-contaminated dust by chicken, resulting in
lead poisoning and teaming disabilities
Cancer Category 1-Assumed no cancer risk since
substances are not currently considered cardnogenlo
Non-cancer Category 4--)nto»riced by widespread
respiratory symptoms and hospital admissions due to
inhalation of elevated add aerosol levels
Cancer Category 3-EstbnatBd 1 to 50 excess annual
in horn inhala
Individual risk
nhalation of a mixture of pollutants; high
Non-cancer Category 3-Ortven by carbon monoxide
exposures, aggravating angina for chronic effects and
causing acute carctoobnpatrment
Cancer Category 3-Estimated 15 to 30 excess annual
cancers from ingeston of fish contaminated with PCBs
NoThcancer Category 4-lnfluenced by ingestion of mercury
contaminated fish at up to 25 tunes the Rf D
Cancer Category 3-Estlmated 20 excess annual cancers
from ingestion of TTHM and arsenic
Non-cancer Category 4-priven by outbreaks of
Glardlarfelated gasirolntettinal dteease
(continued)
m-8
Page 1 of 4
-------
Table 111-3 (Continued)
Relative Residual Public Health Risk Ranking
(un ranked within categories)
ProDlorfi Affea
Substances/Exposuiee
• _ -
Commenu
Category 4 (continued)
Other Ground-Water
Contamination
Pesticide Residues on
Food
Pathogenic Microorganisms.
Sodium. Nitrates. Pesticides.
Solvents, and Petroleum
Products
IHertickte
2 Fungicides
llnsecticidB
1 Growth Regulator
Category 3
Hazardous/Toxic Air
Pollutants
TCE
Carbon Tetrachloride
Asbestos
Discharges to Estuaries.
Coastal Waters, and
Oceans from AD Sources
Asbestos
DDT
HCB
Cancer Category 2-Estimated 1 excess annual cancer from
ingestion of dichloromethane from septic tanks/cesspools
Non-cancer
andt
contamination of drin
4-Driven by outbreaks of hepatitis
caused by septic tank
B-wassr wells
Cancer Category 4-Estimated up to 320 excess annual
cancers by scaling NCRP data down to region, a method that
may overestimate the regional impact
Non-cancer Category 2/1-Assumed tow non-cancer risk
because of lack of evidence that exposures excood allowable
Cancer Category 3-EstimatBd 10 excess annual cancers
based on monitored background concentrations of urban
toxic soup; may underestimate effect because of low
percentage of problem covered
Non-earner Category 2/1-Ouantitativa risk not estimated
because of lack of inhalation reference doses to evaluate
systemic oHucttt from those poDutonts
Cancer Category 3-Estimated 185 excess annual cancers;
work group felt that the estimate was conservative
Non-
Category 2/1-Assumed low non-cancer risk
Cancer Category 3-Estimated 90 excess annual cancers
from ingestion of contaminated fish
Non-cancer Category 2/1-Estimated tow risk since
exposures do not exceed RfDs
Nonpoint Source PCBs
Discharges to Surface Dioxin
Waters Mercury
Cancer Category 3-Estimated less than 30 excess annual
cancers from ingestion of fish contaminated by PCBs; not
judged as significant a contributor as industrial point sources
Non-cancer Category 2/1-Qualitativery estimated as smaller
oonttibutor to surface waters than industtial point sources
(continued)
Page 2 of 4
m-9
-------
Table 111-3 (Continued)
Rotative Residual Public Health Risk Ranking
(unranked within categories)
PrablemAna
Catoooiya
POTWOischanjesto
Surface Waters
mvMtlaated
PCBs
Dtodn
Mercury
Comments
Cancer Category 2-Estimated less than 10 excess annual
cancers from ingestion of fish contaminated with PCBs;
judged a smaller contributor than nonpoht or industrial point
sources
Tanks
Chlorine
Mineral Adds
Organic Solvents
Organic Toxics
Home Healing 03
Diesel Fuel
Fungicides
Rfldtetton from Sourofls
Other than Radon
(non-ionizing)
Extremely Low Frequency
(ELF)Radatton
industrial point sources
Cancer Category 1-Assumedbw cancer risk
i
Non-cancer Category 3-influanoad by respiratory problems
as severe as respiratory pneumonia; estimates based on
reported accident data
Cancer Category 2-QuaHtath» ranking based on best
judgment of risk relative to problem areas with more data;
judged similar to hazardous waste risks
Non-cancer
judgment of risks
3-Quafitative ranking based on best
to other problem areas; high
number of tanks drives higher rating
Cancer Category 2-Estfmated less than 1 excess annual
cancer; derived by scaling down national figures
Non-cancer Category 3-Quaatative ranking based on best
...... to other problem areas
Cancer Category 2-QuaDtative ranking based on best
judgment of risks relative to other problem areas; high degree
of uncertainty in available data
Non-cancer Category 2/1-OuaOtative ranking based on best
judgment of risks relative to other problem areas
(continued)
Page 3 of 4
m-io
-------
Table 111-3 (Continued)
Relative Residual Public Health Risk Ranking
(unranked within categories)
Problem Area Investigated Comments
RCRA Waste Sites 18 Typical Compounds in Cancer Category 2-Estimated 1 to 10 excess annual
Watte Stream cancers. Influenced most heavfly by modeled hazardous
waste incineration impacts; high degree of uncertainty
Non-cancer Category 2/1-Estimated low risk since
exposures do not exceed RfDs
Superfund Waste Sites PCBs. Arsenic, Cancer Category 2-Estimated about one excess annual
ITOflBMlBnpll*WlwlVIIVf
Benzene. Vinyl Chloride
Non-cancer Category 2
exposures generally do not exceed RfDs
Municipal Waste Sites Vinyl Chloride Cancer. Category 2-QuaBtatively estimated a few excess
Arsenic annual cancers due to contamination of ground-water wells;
Tetrechloroetriytene limited available data
Carbon TettachbrkJe Non-cancer. Category-Estimated low risk since
exposures do not exceed RfDs
Industrial Waste Sites Vinyl Chloride Cancer Categoy2-QuaEtaUvetye8timatBd a few excess
Arsenfe armual careers die to contamination of weDs;Dmited
Tetrschloroetnylerie availabtodata
Dtehtoromemane
Carbon Tetrachloflde Non-cancer. Category 2/1-Estimated low risk since
exposures do not exceed RfDs
Page 4 of 4
m-n
-------
Table 111-4
Summary of Relative Residual Public Health Risk Ranking
(unranked within categories)
MghRUt Category 5
Criteria Air PoOutants
Radon
Lead
Category 4
Acid Deposition and Visibility
Indoor Air Pollutants Other than Radon
Industrial Point Source Discharges to Surface Waters
Drinking Water
Other Ground-Water Contamination
Pesticide Residues on Foods
Hazardous/Toxic Air Pollutants
Nonpotnt Source Discharges to Surface Wa
Discharges to Estuaries. Coastal Waters, and Oceans from All Sources
Asbestos
POTW Discharges to Surface Waters
Accidental Releases
Releases from Storage Tanks
Pmttirififi Annlinrtian
Category 1
Radiation from Sources Other than Radon (non-fcnizing)
RCRA Waste Sites
Superfund Waste Sites
Municipal Waste Sites
low W«* Industrial Waste Sites
m-n
-------
IV. Observations
In the "i*u* months from the Public Health Risk Work Group's first meeting to the final
ranking session, we learned much about the relative residual public health risk in the region
and perhaps even more about risk assessment and group decision making. This section
presents the most important of the work group's observations after reflecting on where we
have been and what we have learned. Many observations are not surprising, but they all offer
valuable lessons for those who might venture down a similar path in the future.
• Good data are hard to come by.
The work group ended up analyzing far fewer data than we had originally envisioned. We
attributed this to two major problems. First, data we thought or hoped existed simply did not
exist or existed in such limited quantities that they were of little use. In the second instance,
the data existed in some form but were so decentralized and stored in so many different ways
that it would have taken significant *""g and money to compile them. State agencies possess
vast quantities of valuable data that are often extremely inaccessible. We need to put more
resources into developing and improving data management in Region I and in the states.
• Lack of a standard non
-------
• It is difficult to combine cancer and non-cancer risks.
An enormous amount of personal value judgment is involved in trying to compare and
sum cancer and non-cancer health risks. Although none of the work group members thought
it would be easy, few were prepared for just how uncomfortable and difficult the process
proved to be. Our consensus judgment was well informed and carefully nurtured, but it
cannot be validated or replicated.
IV-2
-------
Appendix
Problem Area Papers
-------
1. Criteria Air Pollutants
Problem Area Definition
This category covers exposure to criteria air pollutants regulated under the Clean Air Act
for the protection of human health and public welfare. The human health effects were
evaluated for the three criteria air pollutants (ozone, paniculate matter less than 10 microns1
and carbon monoxide) that exceed the National Ambient Air Quality Standards (NAAQS) in
Region I. Other criteria air pollutants such as oxides of nitrogen and sulfur were not
included in this evaluation because New England is currently meeting the NAAQS for those
substances, and we simplified our analysis by assuming that exposures below the NAAQS
posed no risk. Lead, the remaining criteria air pollutant, was evaluated as a separate problem
area. Major sources of the pollutants evaluated include motor vehicles, industries, and fuel
burning operations.
Summary/Abstract
This assessment focused on the non-carcinogenic health effects attributed to ozone (63),
paniculate matter less than 10 microns (PMi Q), and carbon monoxide as a result of
exposures to levels in excess of the NAAQS. Carcinogenic effects were not considered in
this assessment because cancer is not thought to be of concern from exposure to these
pollutants. The health effects of concern that served as the basis for the NAAQS and also for
our analysis include increased risk of asthmatic attacks (CM, restricted activity (O* and
PMjg). aggravation of angina (CO), and even premature death (PM10). These effects
represent significant health effects and, as such, the severity scores assigned to these
pollutants were some of the highest of any problem area evaluated. The uncertainty in our
evaluation was low, owing to well-documented health effects and a reliable monitoring
network from which reasonable exposure information was obtained. Because the focus of
the analysis was on health effects in excess of the NAAQS and we could identify the non-
compliance areas within the region, we felt that a large portion of this problem was being
addressed. Health effects associated with levels below the NAAQS were not considered.
^Subsequent to this analysis, the Public Health Risk Work Group decided that there was substantial overlap
between the paniculate matter health effects analysed here and the effects of acid aerosols-one portion of the
paniculate matter dose-analyzed under Acid Deposition and Visibility (problem area #2). We decided to
combine the paniculate matter and acid aerosol health risks to eliminate this overlap. We combined them
under Acid Deposition and Visibility since the acid aerosol analysis indicated the greater public health risk.
Therefore, the particular matter analysis described in this paper was actually considered as pan of the Acid
Deposition and Visibility problem area in ranking the public health risk.
1
-------
Toxicity Assessment
By definition, we were concerned with only those criteria air pollutants for which New
England currently exceeds the NAAQS. Consequently, ozone, PMjQ, and carbon monoxide
were our pollutants of concern. Other criteria air pollutants (oxides of sulfur and nitrogen,
and lead) were not evaluated because Region I is currently meeting the NAAQS.
Because each of the three pollutants evaluated are not currently considered to be
carcinogenic, we focused our evaluation on the systemic effects of these substances. The
health effects of concern are those which served as the basis for EPA's National Ambient Air
Quality Standards. As such, the health effects of concern are as follows: for ozone-asthma
attacks and respiratory restricted activity days; for paniculate matter-premature death and
restricted activity; and for carbon monoxide-aggravation of angina. The National,
Air Quality Standards we focused on are 0.12 ppm (one hour average) for 03, SO ug/nr
(annual average) for PM JQ, and 9 ppm (eight-hour average) for carbon monoxide.
Exposure Assessment
Sources of carbon monoxide and PMin include automobiles, diesel engines, and
combustion sources. Ozone is not emitted directly into our atmosphere but is created by the
interaction of sunlight with hydrocarbons and NOX. Thus, many of the large industrial
facilities that emit volatile organic compounds, combustion sources, cars, and trucks emit
The only pathway of exposure considered for this problem area was through direct
inhalation.
Information from the 1986 Annual Air Quality Report for New England was used to
obtain data on the number, location, and actual pollutant levels in excess of the NAAQS
from monitors located throughout New England. These data were used with probabilistic
equations developed by a consultant for Region X to estimate the actual incidence of
different health effects attributable to ozone and PMjg. Because a probabilistic equation
was not available for carbon monoxide, we made a crude estimate of the number of people
exposed to levels in excess of the eight-hour standard and attempted to characterize the
population at increased risk of angina.
The population at risk of experiencing an asthma attack as a result of exposure to ozone
levels in excess of the standard are those who experience levels in excess of the standard and
who are asthmatics. Because asthma attacks may be brought on by acute exposures to ozone
-------
occuning repeatedly throughout the course of a year, we built in a measure of exposure
frequency into our population estimate. Thus, for ozone, the population at risk translates to
approximately 2.8 million person-days based upon estimates of the population exposed to
ozone levels in excess of the standard (72 million person-days) and the prevalence of
asthmatics in the population (approximately 4 percent). The remaining 70 million
person-days in Region I were evaluated as being at risk of experiencing restricted activity on
days when ozone levels exceeded the standard.
For paniculate matter, only one monitoring location in all of Region I exceeded the
annual average. Thus the population in this city (New Haven, CT--population 120,000) was
deemed at increased risk of premature death and restricted activity.
For carbon monoxide, eight urban areas with a total population of 1.2 million people
were deemed at risk of exposure to carbon monoxide levels in excess of the eight-hour
standard. Of these people, approximately 96,000 were deemed to be at increased risk of
having their angina aggravated as a result of exposure to carbon monoxide based on the
prevalence of angina (8 percent) in the U.S. population.
Risk Characterization
Cancer risk estimates were not performed because the pollutants evaluated are not
thought to be known or suspect human carcinogens.
The general approach we used to assess the non-cancer health effects of criteria air
pollutants was to generate estimates of annual incidence. To do this we relied on several
equations developed by a contractor working on this problem for another region. The
equations predicted incidence given exposure information, the population, and pollutant
levels. This was done for both ozone and PMi Q because we had the incidence equations
developed for Region X's Comparative Risk Project. For carbon monoxide, an incidence
equation was not available; thus a simple ratio of the exposure levels to the eight-hour
standard was used to describe potency. The results of our calculations follow:
• Ozone
For ozone, two health endpoints were evaluated: increased asthmatic attacks and
restricted activity days. These two endpoints were both given a moderate severity ranking (2
on a scale of 1 to 4). The incidence values computed using Region X's equation together
with Region I exposure information resulted in annual incidence estimates of approximately
3.7E-4 and 1.2E-3 for asthmatic attacks and for restricted activity days, respectively. This
incidence information (substitute for potency estimate) justified issuing a high potency score
to each of these endpoints (3 on a scale of 1 to 4). Because the populations for both
endpoints were very large (in excess of one million), the highest population score (4) was
assigned to each.
-------
• Particulate Matter
Premature deaths were estimated based on exceedances of the annual average standard
and a probability (incidence) equation developed by a contractor for Region X. Possible
premature deaths were considered extremely severe and thus warranted the highest severity
of effect ranking score. The predicted incidence of death given Region I exposure
information was 5E-6, which translated to a mild potency value of 2. The population at risk
was assumed to be the entire urban population (120,000) of the city in which the annual
average was exceeded, which placed this endpoint at the high end (population score of 3).
Restricted activity (severity score 2) was also evaluated. The predicted incidence was much
higher (4E-2), warranting a potency score of 4. The population at risk was the same
population as above, namely the 120,000 urban dwellers, corresponding to a population
score of 3.
• Carbon Monoxide
Since incidence data were not readily available and since we did not have any means to
estimate annual incidence, the average CO levels in the CO non-attainment areas were
compared to the NAAQS to compute a potency score. Since the levels in exceedance areas
were only slightly above the eight-hour standard, the potency score assigned to CO was low
(1). The health effect of concern at the standard is aggravation of angina, which warranted a
severity score of 3. The population estimates were based on the number of people who have
heart conditions in areas currently exceeding the NAAQS. This was estimated to be
approximately 8 percent of the urban population from several different urban areas, or
approximately 96,000 individuals, warranting a population score of 3.
Because cancer effects were not evaluated for this problem area, only the non-cancer
effects had to be combined. To accomplish this, each of the severity, potency, and
population scores were summed for each pollutant evaluated.
Uncertainty
The uncertainty in this analysis is attributed predominantly to the exposure analysis; the
hazard identification of acute effects has been studied extensively and is well documented.
We only considered effects associated with exposure to levels above the current NAAQS.
Thus, we may have overlooked health effects associated with exposure to levels below the
standard. The 1986 Ambient Air Quality Report for New England served as our source of
data on documented exceedances. Although we can be very confident of the monitoring
data, it is limited to a few locations in each state. As a result, the areal extent of actual
violations of the NAAQS may be underrepresented. For these reasons, the risks associated
with this problem area may be underestimated.
Estimates of the number of people in the vicinity of each monitor (NOX, CO) were crude
at best, and for ozone we assumed mat entire urban populations were exposed. Estimates of
the exposure to ozone in excess of the standard, reported in terms of person-days, were
generated by staff in the Air Management Division.
-------
Because incidence estimates were generated using an equation which did not receive
much scrutiny from the Regional Office, the incidence estimates for the health effects
associated with ozone and PMjQ are subject to some uncertainty. However, the uncertainty
assigned to this problem area was deemed to be low in comparison to the other problem
areas.
We felt that the percentage of problem covered in this analysis was high even though the
focus was only on three pollutants. This is because Region I is currently in attainment for
SOX,NOX, and lead.
-------
2. Acid Deposition and Visibility
Problem Area Definition
This problem area concerns health effects associated with acidic aerosols and acid
deposition. Some gases emitted into the atmosphere react with sunlight, water vapor, and
oxygen to form acid compounds. When these compounds are suspended in the atmosphere,
they are referred to as acidic aerosols. Acid deposition occurs when these compounds
interact with precipitation or with small particles to deposit on land and water surfaces.
The primary precursor pollutants that lead to acidic aerosol formation are sulfur dioxide
and nitrogen dioxide. Sources of these primary pollutants include electric utilities (especially
coal-fired), industrial boilers and heaters, smelters, residential and commercial fuel burning,
and mobile sources. Emissions of volatile organic compounds also contribute to the
formation of acidic aerosols.
Most of the acidic aerosols that form in the atmosphere consist of very fine acid sulfate
and nitrate particles. The health effects of primary concern here are associated with the
inhalation of sulfate aerosols. Other possible effects, associated with the leaching of toxic
metals into drinking water, will be evaluated in the drinking water problem area.
Summary/Abstract
The risks associated with this problem area include respiratory symptoms, hospital
admissions, and possibly elevated mortality rates due to the inhalation of acidic aerosols. The
population exposed is large (more than 1 million people), but the population at risk for some
endpoints of concern is much smaller. The incidence-to-population ratios are fairly high
(about 3E-2 for respiratory symptoms and 2E-S for death). Hospital admissions due to
respiratory illness were calculated to pose the highest risk for the endpoints analysed, with a
potency score of 3, a population score of 4, and a severity score of 3. (The maximum score
for any of the three factors was 4.) The percentage of problem covered was considered high,
and the uncertainty was considered high as well.
^Subsequent to this analysis, the Public Health Risk Work Group decided that there was substantial overlap
between the effects of paniculate matter analyzed under Criteria Air Pollutants (problem area #1) and the
effects of acid aerosols-one portion of the paniculate matter dose-analyzed here. We decided to combine
paniculate matter and acid aerosol health risks to eliminate this overlap. We combined them under Acid
Deposition and Visibility because the acid aerosol analysis indicated the greater public health risk. Therefore,
the paniculate matter analysis described in the Criteria Air Pollutants paper was actually considered as part of
the Acid Deposition and Visibility problem area in ranking the public health risk.
-------
Toxicity Assessment
Evidence suggests that exposures to acidic aerosols are associated with respiratory
symptoms, such as cough, in adults and children with chronic respiratory diseases as well as
in healthy individuals. Associations have been observed between sulfate levels and mortality
rates and hospital admissions due to respiratory illnesses. Laboratory work with animals
suggests that acidic aerosols could be related to the development of chronic respiratory
diseases, but this has not been confirmed with human epidemiological studies.
We have estimated four categories of health effects (deaths, hospital admissions,
respiratory symptoms in children, and respiratory symptoms in adults) using the following
equations recommended by our expert consultant, RCG/Hagler, Bailly, Inc.:
Annual deaths = .000037 * (Sj -10) * POPj
where: Sj = annual average sulfate level in area j (ug/cubic meter)
POPj = population in area j
Annual admissions = .000086 * Sj/9 * POPj
Children with respiratory symptoms = .035 * C
where: C = the number of children in areas where the annual average sulfate level
exceeds 10 ug/cubic meter
Adults with respiratory symptoms = .05 * M + .02 * W
where: M = the number of men in areas where the annual average sulfate level
exceeds 10 ug/cubic meter
W = the number of women in areas where the annual average sulfate level
exceeds 10 ug/cubic meter
The work group assigned the following severity scores to the health endpoints of concern for
the non-cancer health effects analysis:
deaths: 4
hospital admissions: 3
respiratory symptoms: 1
Exposure Assessment
For the exposure assessment, we used the available monitoring data on ambient sulfate
levels in the region as an indicator of the population exposure to potentially harmful acidic
aerosols. Some data points were eliminated from further consideration in order to exclude a
few local source-dominated impacts. The group felt it was not reasonable to associate these
-------
hot-spot impacts with larger populations that are probably exposed to lower levels of sulfates.
In addition, some older data were eliminated because the group felt they were no longer
representative of current ambient levels.
In most cases where a city had more than one sulfate monitor, monitor readings were
averaged to obtain a single characteristic value for the city. Sulfate levels in a city were
associated with the city population. Rural monitor levels within a county were averaged, and
the resulting average concentration was associated with the county population. If the county
included cities for which monitoring data were available, then the rural monitor levels were
associated only with the rural population in the county. Sulfate levels were not estimated for
counties without monitors.
i
The number of individuals in New England exposed to acidic aerosols is greater than
one million, which translates to a population score of 4 in our ranking system. However, the
number of individuals at risk for some of the endpoints of concern (those exposed to annual
average levels greater than 10 ug/cubic meter for death and respiratory symptoms) is much
lower.
Risk Characterization
A small number of excess annual deaths was predicted using the methodology suggested
by the consultant The analysis for deaths produced an estimated incidence/population ratio
of 2E-5 for the population at risk. This corresponds to a potency score of 2 in our scoring
system, with a population score of 3 and a severity score of 4.
The analysis for increased hospital admissions due to respiratory illnesses produced an
incidence/population ratio of 1.2E-5 for the population at risk. This represents a potency
score of 3, with a population score of 4 and a severity score of 3.
The methodology recommended by the consultant for estimating the increased incidence
of respiratory symptoms in children implicitly uses an incidence/population ratio of 3.SE-S for
the population at risk. The analysis for the increased incidence of respiratory symptoms in
adults produced an incidence/population ratio of approximately 2.6E-6. These represent
potency scores of 4, with a population score of 3 and a severity score of 1.
The analytic procedures used to estimate the incidence of health effects are based on
epidemiological study results that demonstrate associations, but not causation. Therefore,
there is some uncertainty regarding the magnitude and possibly the existence of the assumed
causal relationships used here to relate pollutant levels to human health effects.
The equations recommended by the consultant to estimate excess mortality and increased
respiratory symptoms make use of a threshold value of 10 ug/cubic meter. The selection of
this specific threshold value is somewhat arbitrary, since the available scientific information
does not allow for a more precise estimation of a suitable threshold.
-------
The results of the exposure and risk analyses are sensitive to how multiple data points in
cities are treated and to what populations are selected to associate with monitored levels. Very
few recent monitor sites exhibited annual sulfate levels greater than 10 ug/cubic meter. One
site in Maine was not carried through the full analysis because the work group believed that
the monitored levels represented elevated impacts from a local source and were not indicative
of the type of problem for which this problem area was created. The remaining monitor site
with recent annual sulfate levels greater than 10 ug/cubic meter was in Springfield, Mass.
Another site in Springfield exhibited annual levels somewhat lower. The work group decided
to associate the higher concentration measurement with the population of Springfield, rather
than averaging the levels at both monitors. Averaging the two monitored levels would have
produced an average concentration below 10 ug/cubic meter and eliminated the excess
mortality and increased respiratory health effects entirely.
In these analyses, we relied upon the use of available information on ambient sulfate
levels as an indicator of population exposures to potentially harmful acidic aerosols. It must
be noted that certain types of sulfates, such as sulfuric acid, may be more harmful than other
types of sulfates and that acidic nitrates may also represent a potential health hazard. The
proportion of the total problem represented by sulfates would be expected to vary, and
therefore it is difficult to specify what portion of the harmful acidic aerosol total is represented
by sulfates. Our consultant thought that sulfates represented much more of the problem than
nitrates.
Another factor to consider is the correlation between sulfate levels and levels of total
particulates. Attributing the health effects of concern here entirely to acidic aerosols may
somewhat overestimate the degree of their culpability.
The uncertainties associated with the use of ambient sulfate levels as an indicator and with
the correlation between sulfate and total paniculate levels may be reduced when more
monitoring data for a variety of acidic aerosols become available, and when more
epidemiological studies are able to make use of such data.
An important potential health effect that was not considered in the analysis is the
development of chronic respiratory disease. Although it may be reflected to some extent in
the estimates of other health effects that were considered, chronic respiratory disease may
represent another important health problem associated with the long-term exposure to elevated
acidic aerosol levels.
The combined effect of these uncertainties is not clear. To the extent that some health
effects due to other pollutants are being attributed to acidic aerosols, these estimates may be
too high. Conversely, to the extent that significant health effects are not considered and to the
extent that other acidic aerosols are not accounted for in the analysis, these estimates may be
too low.
The work group considered the percentage of the problem covered to be high and the
uncertainty to be high as well.
-------
3. Hazardous/Toxic Air Pollutants
Problem Area Definition
This problem area evaluates risks from ambient exposure to hazardous air pollutants. This
is defined to include a very broad set of substances that currently are not directly regulated
under the Clean Air Act, although they may indirectly be controlled as components of
regulated mixtures such as volatile organic compounds. Monitoring data for chromium,
arsenic, benzene, perchloroethylene, trichloroethylene, and carbon tetrachloride were
considered. Because monitoring data are so limited and poorly organized, modeling data
were also evaluated for a broader range of pollutants.
Sources of these pollutants include point sources such as incinerators, fugitive emissions
from solvent use, and area sources such as motor vehicle emissions.
Summary/Abstract
Individual risks based on monitoring data and professional judgment were estimated to be
in the 1E-4 range for lifetime exposure. The population exposed to this risk was estimated to
be 3.4 million. Data provided by a modeling analysis predicted higher risks. Risk estimates
were not considered from "high risk point sources" but only from background urban
exposures. Non-cancer risk estimates were also not considered because of the lack of
inhalation Reference Doses to evaluate systemic effects from these pollutants.
The data used to make these risk estimates have a large range of uncertainty due to the
lack of routine monitoring for these pollutants, the lack of standard methods for such
monitoring, and an inability to predict how well ambient air data serve as an estimate of actual
human exposure.
Toxicity Assessment
Because of the broad range of pollutants considered, it is not possible to present detailed
discussions of the toxicity of these pollutants. Attachment 3- A shows unit risk estimates for
the six pollutants evaluated based on monitoring data. These include two metals and four
YOCs. The metals have unit risk estimates [ug/nr]'1 in the 1E-2 to 1E-3 range, while the
VOCs have unit risk estimates in the 1E-6 range.
• Hazard assessments for these chemicals are based on data of widely varying quality and
certainty. Arsenic, chromium, and benzene are considered by EPA to be class A carcinogens,
indicating that there are human data available to support the conclusion that they are
carcinogens. There is a major uncertainty about chromium, however, since it is considered
that only Cr+6 and not Cr+3 is carcinogenic.
10
-------
Data on many VOCs are based on animal bioassays. There is considerable controversy
about many of these VOCs, such as perchloroethylene, over what their hazard classification
should be. Perchloroethylene, in particular, is now somewhere between a "C" (possible
human carcinogen) and a "B2" (probable human carcinogen) in the agency's ranking. These
uncertainties, which tend to get lost in quantitative evaluations of risk, need to be considered.
In summary, a small number of these pollutants are almost certainly human carcinogens while
there is great uncertainty about the rest.
There is even greater uncertainty concerning non-carcinogenic or systemic effects of these
pollutants. Other than the criteria pollutants such as lead or paniculate matter, the air program
has not evaluated systemic effects for air pollutants in any systematic way. This is in contrast
to EPA's evaluation of oral exposure where "risk reference doses" (RfDs) are well established
for a large number of pollutants. There is a work group developing inhalation RfDs within
the agency, but its methods are not yet final.
Other approaches, such as use of occupational standards divided by safety factors, have
also been used, particularly by state air programs. These were not considered useful for this
analysis, however. As a consequence, only cancer risks are considered, using standard CAG
unit risk estimates. These estimates assume lifetime exposure at a breathing rate of 20 cubic
meters per day.
Exposure Assessment
The most difficult aspect of this analysis was the compilation and evaluation of exposure
data. This difficulty can be attributed to the lack of a regulatory structure to control these
pollutants. Because there is no federal regulatory program, sources of monitoring data are
limited to a small number of research programs and special studies. Li addition, the data that
are available have not been compiled in a systematic way that would allow a comprehensive
evaluation of risks. Finally, even where data are available, there are major uncertainties
caused by the lack of standard air toxics monitoring methods.
There were, however, some data that were evaluated from the following sources:
• EPA's Toxics Air Monitoring System (TAMS): Chelsea, MA; Boston, MA (VOCs)
• EPA's NMOC Program: Portland, ME; Boston, MA; New Haven, CT; and Bridgeport,
CT(VOCs)
• EPA's National Paniculate Filter Analysis Program: 25 cities (metals)
• A report from the MA air toxics program (VOCs)
11
-------
Data from the first three of these sources were entered into a Lotus spreadsheet. The
Massachusetts study was reviewed and compared with data from the other sources. The
standard averaging time for these studies was 24 hours. Although several days of
observations were available in most cases, extrapolating from a small number of 24-hour
observations to annual levels introduces another level of uncertainty.
The metals data showed large variations in concentrations for arsenic, cadmium, nickel,
and chromium. Of these four, the data for arsenic and chromium appeared to be the most
consistent and were used to develop the risk estimates discussed later. A summary of this data
is presented in Table 3-1, along with risk estimates.
A major risk issue with chromium concerns the percentage that is Cr+6 as opposed to
Cr+3, because only Cr+6 is considered to be carcinogenic. For the risk estimates developed
here, it was assumed that one-seventh of the total Cr was Cr+6. This ratio had been used in an
epidemiological study of Cr health effects, but could have a large margin of error.
The VOC data were much more limited in geographic scope, but for the TAMS and
NMOC studies analyses had been done for dozens of organic compounds. Four VOCs were
used to develop risk estimates: benzene, perchloroethylene, trichloroethylene, and carbon
tetrachloride. These were chosen for their prevalence and for the availability of unit risk
estimates. A number of other VOCs, such as 1,1,1-trichloroethane, 1,1-dichloroethene, and
various aromatics, were omitted. The data are summarized in Table 3-1.
No ambient data were found for formaldehyde, which in other studies has been shown to
contribute to a significant fraction of the overall risk from ambient exposures.
It should be noted that all of these data pertain to background urban exposures ("urban
soup"). Sources of such exposures include motor vehicle emissions, combustion of fuel in
boilers, solvent use by dry cleaners and other commercial establishments, and emissions from
gasoline stations.
It does not address levels in non-urban areas (which are likely to be lower), and it also
does not address exposure from "high risk point sources" around which levels of these and
other pollutants are likely to be much higher. Such high risk sources include chemical
production plants, combustion units such as waste incinerators, and processors of large
amounts of solvents.
In addition to this monitoring data, modeling data were developed by a consultant (but
completed after the work group ranking). These data were developed from national emission
estimates for a large number of pollutants. The advantage of these data are that they are much
more comprehensive than the monitoring data in geographic scope and in number of
pollutants covered The disadvantage is that the emission estimates on which the modeling is
based have a large potential margin of error. In addition, uncertainty is introduced because of
the modeling.
12
-------
Modeled data were available for 22 pollutants. A summary of this data is provided in
Table 3-2. As can be seen from comparing Tables 3-1 and 3-2, the modeling data predict
much higher risks than the available monitoring data. This in part is due to more conservative
assumptions used in the modeling (such as 24 hours per day exposure versus 8 hours per day
assumed for the evaluation of the monitoring data) but may also be due to the tendency of the
emission estimates to overstate actual emissions. At this point, the only statement that can be
made is that there is a discrepancy which warrants further evaluation.
It is likely that the monitoring data set an upper bound on the risk levels from ambient
exposures to hazardous air pollutants, not considering the exposure to "high risk point
sources".
Risk Characterization
Risk has been characterized only for the carcinogenic risks of hazardous air pollutants. As
noted above, standard methods are not available for developing inhalation risk reference
doses, so non-cancer effects have not been considered. There would also be some potential
double-counting with the criteria pollutant evaluation. It should be realized, however, that in
addition to the carcinogenic effects discussed below, these pollutants could also be present in
amounts large enough to cause concern about systemic effects.
Table 3-1 summarizes the risk evaluation based on monitoring data. Using mean values
of these data, individual risks for each pollutant are generally in the 1E-6 range, while
cumulative risks are in the 1E-5 range. The evaluation assumes eight hours per day exposure,
and assumes that Cr+6 is one-seventh of total Cr. The evaluation does not include a large
number of pollutants for which ambient data were not available. It was, therefore, the
consensus of the work group that the actual risk of the total mixture of pollutants (including
such substances as formaldehyde and BAP) was likely to be in the 1E-4 range.
The population to which this risk applies is considered to be those living in urban
areas—about 3.4 million, based on 1980 census data. In addition, there are transient
populations, such as commuters, which are exposed to these pollutants for a shorter length of
time. Using the 3.4 million population figure, lifetime risks are in the range of 70 (for the
6 pollutants considered) to 700 (assuming a 10-fold increase in risk when other pollutants are
added).
As noted above, the modeling data predict higher lifetime risks. As shown in Table 3-2,
best-case lifetime population risk for all pollutants is about 1,800 cancer cases, while
worst-case estimates are about 8,300. The worst-case estimate is about an order of magnitude
higher than for the monitoring data risk assessment while the best-case estimate is about a
two-fold increase over the monitoring estimates.
13
-------
Table 3-1
Summary of Air Toxics Monitoring Data
Pollutant
ChromIum(+6)
Arsenic
Benzene
Perchloroethylene
Trichloroethylene
Carbon Tetrachloride
Number
of Sites
25
25
2
2
1
1
Maximum
0.0659
0.0110
3.0200
295.7000
1.9600
1.7800
(UG/L)
Minimum
0.0000
0.0000
0.0000
0.0000
1.0500
0.8900
Mean
0.0049
0.0024
1.6700
16.4900
1.4900
1.3400
Unit Risk
(ug/m3)'1
1.02E-02
4.30E-03
8.03E-06
5.80E-07
1.30E-06
1.50E-05
Individual
Risk
2.38E-06
3.44E-06
4.47E-06
3.19E-06
6.46E-07
6.70E-06
Urban Total
Total Population Risk
2.08E-05 3.40E+06 70.80
Notes: Assumes Cr+6=1/7 of total Cr.
Assumes exposure ° 8/24 hours/day.
14
-------
Pollutant
Table 3-2
Summary of Air Toxics Modeling Data
Lifetime Population Cancer Risk
Best Case
Worst Case
Arsenic
BAP
Benzene
Beryllium
1,3 Butadiene
Formaldehyde
EDB
Cadmium
Chloroform
Carbon Tetrachloride
Gasoline Vapors
Cr+6
Methylene Cl
PCB
PERC
POM
TCE
Vinylldene Cl
8.3E+01
1.1E+01
1.7E+02
1.5E+00
3.6E+00
3.9E+02
4.8E+01
1.2E+01
9.8E+00
2.7E+01
1.1E+00
8.8E+02
9.0E+01
5.8E+00
6.9E+00
1.4E+00
4.5E+00
6.1E+01
4.0E+02
7.0E+01
9.7E+02
8.0E+00
2.9E+02
1.7E+03
2.2E+02
6.7E+01
5.2E+01
1.1E+02
8.9E+01
3.5E+03
5.1E+02
2.3E+01
5.7E+01
3.9E+01
4.7E401
2.4E+02
1,806
8,392
15
-------
4. Radon
Problem Area Definition
This problem area covers risks from exposure to indoor radon through all routes of
exposure, including inhalation of indoor air and exposure to radon gas released from drinking
water. Only indoor exposures are considered.
Summary/Abstract
Radon is a naturally occurring inert radioactive gas. Radon 222 is produced as a decay
product of Uranium 238, which is found naturally in certain rock types, and has a half life of
3.8 days. Radon decays into other short-lived radioactive progeny which can be inhaled and
adsorbed within the lungs. Sources of radon include rocks and soil around a home and
underground water from uranium-bearing aquifers, although migration of radon through soil
pore space and into a home is considered to be the major source of indoor exposure.
An increased risk of developing lung cancer is the health effect of concern resulting from
exposure to elevated levels of radon. Nationally, EPA estimates that 5,000 to 20,000 lung
cancer deaths per year can be attributed to radon. These risk estimates are based on
epidemiological studies of miners, and employ use of a "relative risk" model that ties the risk
of lung cancer due to radon to the overall rate of lung cancer in the country. Risks of lung
cancer from combined exposure to radon and smoking are greater than the sum of risks from
exposure to either one alone.
Region I has maintained a database of indoor radon measurements that was used to
develop regional risk estimates for this study. These estimates were based on more than 4,000
basement measurements taken on both a random and non-random basis in all Region I states.
Using the risk equation that the Office of Radiation Programs (ORP) developed to estimate
national risks, it was estimated that 375 to 1,500 lung cancer deaths per year could be
attributed to radon exposure. An additional analysis of annual cancer incidence attributable to
ingestion and inhalation of radon from drinking water was conducted (see page 20). It is
estimated that approximately 75 additional annual cases could be attributed to this route of
exposure, for a total range of 450 to 1,575 cancers per year.
Toxicity Assessment
As stated above, risks from radon are due to exposure to decay products. Collectively,
these decay products are measured in units of working levels (WL), although radon itself is
usually measured in picocuries per liter (pCi/L). In order to develop risk estimates based on
radon concentration data, it is necessary to convert from pCi/L to WL, and, for this analysis, a
standard conversion factor of 0.005 is used.
-------
The equation ORP used to develop national radon risk estimates is shown in
Attachment 4-A. Elements of this equation are discussed briefly here. A more detailed
discussion can be found in EPA's Radon Reference Manual, EPA S20/1-87-20. The first
element (CR) is the average indoor radon decay product concentration that, for the national
estimate, was one average for the whole country. As discussed below, the regional estimate
was based on a range of radon measurements broken into percentiles.
The second element (T) is the time (in hours) of exposure to radon which is assumed. The
ORP equation assumes lifetime exposure, but also assumes 75 percent occupancy in the home
and subtracts a 10-year latency period. The third factor (F^yiM) converts WL hours to WL
months, which is the unit for which risks to miners has been developed. This factor also takes
into account the difference in breathing rates between a miner and the average adult.
The next factor (RRRM) is the estimate of relative risk per WL month. This factor is
based on the relative nskmodel which assumes that the risk of developing a tumor because of
radon exposure is relative to the underlying lung cancer rate in the overall population. The
relative risk model projects an expected percentage increase in lung cancer rates. The
percentage is based on epidemiological studies of lung cancer rates in miners. About 13 such
studies have been conducted. There is a range of relative risks that can be developed from
these studies, and the range that ORP uses (which has been supported by the Science
Advisory Board) is 1 percent to 4 percent
The next factor is the underlying annual average lung cancer rate (TCR), which (from
1980 vital statistics) is 4.584E-4 per person. The last factor (POP) is the population to which
the risk estimates are applied. For the national estimate, the total U.S. population was used.
For the regional analysis, an estimate of the population living in detached structures was used.
Exposure Assessment
Low levels of radon can be found everywhere. In the indoor environment, exposure to
radon mainly results from radon in rocks and soil containing uranium or radium. Radon gas
can then enter a home through floors or walls as a result of either diffusion or pressure driven
flow. Radon can also enter a home through ground water that contains radon as a result of the
rock formation from which it comes. The radon is then released when water is used in the
home.
The regional estimate of radon risk was developed based on data taken from more than
4,000 New England homes. These data come from a variety of sources including random
studies of homes in CT and RI and non-random data supplied by radon testing companies in
the other states. All of the data used were from tests done in basements where radon levels are
expected to be highest These tests are done using charcoal canisters which serve as a
short-term screening method to detect radon levels during a 24- to 48-hour period.
17
-------
Attachment 4-B summarizes these data. For this analysis, the measurements were broken
into more than 80 fractions based on the range of radon concentration measured. For each
fraction, the percentage of total measurements in that fraction was determined. (For example,
about 22 percent of homes had radon levels between 1 and 2 pCi/L). The measurements in
pCi/L were then converted to WL.
The next step was to determine the population in each fraction. To do this, 1980 census
data were used. Since the measurements and the associated risk apply to persons living in
single family homes or in other one-story dwellings (such as duplexes), an estimate of that
population was made by taking the number of single family homes (2.49 million) and
multiplying it by the average number of persons per household (2.74) to get an exposed
population of about 6.8 million.
An estimate of persons living on the ground floor of multi-unit buildings was added to this
number by taking the number of persons living in households with 2 to 4 units (about
3 million) and assuming that 50 percent of persons living in those households lived on the
ground floor (about 1.S million). Adding this to the previous figure gives a total estimate of
about 8.3 million for the population at risk.
It was then assumed that this population was exposed to the same distribution of radon
concentrations as was the sample in the regional data base. The population was then
distributed according to the distribution of measurements in that data base. Basement
measurements were converted to an estimate of radon levels in living areas by multiplying by
a correction factor of 0.3. The ORP radon risk equation was then used to estimate a low
(assuming a 1 percent relative risk) and a high (assuming a 4 percent relative risk) risk
estimate.
Risk Characterization
A shown in Attachment 4-B, the evaluation estimates between 375 and 1,500 annual lung
cancer deaths that can be attributed to radon. This corresponds to an annual individual risk
for the exposed population of about 4.5E-5 to 1.8E-4.
The analysis has a number of strengths and uncertainties. First, the estimates of radon risk
are based on human data, in contrast to many environmental pollutants for which toxicity is
evaluated from data derived from animal studies. In addition, the range of radon
concentrations found in homes is not, in many cases, different from the concentrations found
in the mines where the studies were done. Therefore, there is less of a problem extrapolating
from high to low doses. Finally, the exposure data are based on a relatively large number of
samples taken within homes. This provides the basis for some confidence that the
concentrations evaluated in this risk evaluation are representative of levels to which people are
actually exposed.
18
-------
There are also many uncertainties in the evaluation. The relative risk model assumes that
risks from radon are tied into the overall lung cancer rate. Most of this lung cancer can be
attributed to smoking, and there is uncertainty as to how much of the risks from radon
exposure are due to combined radon/smoking effects. ORP has attempted to estimate what
percentages of the total lung cancer deaths are due to smoking, radon, and smoking/radon.
These data show that if smoking were eliminated the number of cases caused in part by radon
would decrease greatly.
There are other uncertainties in the risk estimates, such as the quality of the miner data and
the extrapolation from studies of healthy male workers to the general population. It should be
noted that the National Academy of Sciences recently evaluated radon risks and came up with
estimates that were close to ORP's.
There is also uncertainty in the exposure estimates due to the extrapolation from basement
to living area concentrations, in the estimate of the population exposed, and in the assumption
that persons are exposed throughout their lifetimes to the radon levels mat were evaluated.
The percentage of problem covered is high. However, it should be noted that radon
exposure through ground water was not explicitly evaluated. The radon data base may have
missed this route of exposure since basement measurements are designed to capture radon
exposure through rock and soil emissions and not through household water use.
Radon in Drinking Water
Summary
Individual cancer risk derived from exposure to radon in drinking water is about 8.3E-4.
The annual cancer incidence attributable to ingestion and inhalation of radon from drinking
water is approximately 75 cases.
Hazard Assessment
Radon is a natural occurring radionuclide and is a human carcinogen (Group A). The
Cancer Potency Factor (CPF, or slope factor) is 2E-8 to 6E-6 per pCi/L. A CPF of 5E-7 per
pCi/L of radon in water was used in this analysis.
19
-------
Exposure Assessment
The radon data were based on the following: Radon in Vermont Public Water Supplies
(1986), A Survey ofRadon-222 Occurrence in Connecticut Private Well Water; Assessing
Geologic and Hydrologic Parameters (1987), and the monitoring data compiled by the Water
Supply and Ground Water Management Branch in Region I.
The median concentration of radon in drinking water was estimated to be 1,745 pCi/L.
The exposed population was estimated to be 6 million ground-water users in Region I.
Routes of exposure considered in this evaluation were ingestion and inhalation of the
radon emitted into the air as a result of all household water supply uses. Exposure from
dermal contact is considered minimal.
Risk Characterization
The individual cancer risk is estimated at 8.3E-4. The annual cancer incidence attributable
to radon exposure from drinking water is approximately 75 cases.
Uncertainties
Radon data from various sources are used. The validity of the data and their
representation are difficult to evaluate. The CPF estimated for radon is 2E-8 to 6E-6 per
pCi/L of radon in water. A CPF of 5E-7 per pCi/L for radon is chosen to estimate the cancer
risk derived from exposure to this compound from drinking water. It is uncertain whether the
selection of 5E-7 per pCi/L is an underestimate or an overestimate.
20
-------
Attachment 4- A
Simplified Approximation of Estimates For
Total U.S. Lung Cancer Deaths Due to Indoor Radon
A. Equation
Total U.S. Lung Cancer Death From
Indoor Radon (1980) = CR * T * Vy/iM * RRRM * TCR * POP
where:
CD = average (mean) lifetime indoor radon decay product concentration
= 0.004 WL-life
T = average interval of lifetime exposure in hours, following a 10-year minimum
induction period during which no lung cancer will be observed, assuming
75% occupancy and 73.88 years life span (1980 vital statistics)
= 0.75 * (73.88-10) * 365 * 24 = 419,691.6 hours/life
FWLM = fe0*0? converting average cumulative indoor exposure in WL hours to
working level months (WLM) for a miner (since risk estimates are based on
miner data), accounting for 170 hours per month exposure period per WLM
(by definition), and the difference in breathing rate between the average
adult (15.3 liters per minute) and a miner (30 liters per minute)
= 1/170 * 15.3/30 = 0.003 WLM per hour
relative lung cancer risk for lifetime exposure to radon, per WLM, using
relative risk model
= 1% to 4% per WLM
TCR = underlying annual average of U.S. lifetime lung cancer, risk (1980 vital
statistics)
= 4.584 * 10"4 per person
POP = 1980 U.S. population
=226,545,805
21
-------
B. Calculation
Total Lung Cancer Deaths
Total Lung Cancer Deaths
0.004 * 419,691.6 * 0.003
* 0.01 (lower risk estimate)
* 4.584 MO'4* 226,545,805
5.230
0.004 * 419,691.6 * 0.003
* 0.04 (upper risk estimate)
* 4.584 * 1CT4 * 226,545,805
20.921
C. Notes
The above calculations differ from the estimates in "A Citizens Guide" of 5,000 to 20,000
lung cancer deaths principally due to two simplifications in the equation used above:
• The factor FWLM does not include correction for the smaller lung size and lower
breathing rate oicnildren. Both are, in fact, recognized in the detailed analysis.
• The product of TCR and FOP is replaced in the detailed analysis by calculations using
1980 mortality rates and 1980 life table statistics. The detailed actuarial analysis more
properly accounts for latency effects, competing risks, and the lower underlying risk of
lung cancer at young ages and, hence, results in a lower estimate.
Use of the 10-year latent period leaves an average life span of 63.88 years (73.88 years-10
years) during which the potential excess lung cancer risk can be expressed.
Source: Putnam, Hayes & Bartlett, Inc., September 1987. The numerical assumptions
listed in Part A of this table are based on EPA analysis. Specific sources are
noted in the text
22
-------
Approximate Yearly Lung Cancer Deaths from
Exposure to Airborne Radon In New England
pCI/L Fraction
Range In Range
1 0.287394
2 0.223230
3 0.141530
4 0.080049
5 0.060037
6 0.037961
7 0.026820
8 0.020012
9 0.017949
10 0.016711
11 0.008665
12 0.008046
13 0.008665
14 0.004332
15 0.007839
16 0.002063
17 0.005364
18 0.003713
19 0.002269
20 0.002475
21 0.002682
22 0.002475
23 0.001444
24 0.001237
25 0.001237
26 0.001650
27 0.001650
28 0.001856
29 0.001031
30 0.001031
31 0.000618
32 0.000412
33 0.000206
34 0.000825
35 0.000412
36 0.001031
37 0.000825
38 0.000206
WL
Range
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
0.045
0.05
0.055
0.06
0.065
0.07
0.075
0.08
0.085
0.09
0.095
0.1
0.105
0.11
0.115
0.12
0.125
0.13
0.135
0.14
0.145
0.15
0.155
0.16
0.165
0.17
0.175
0.18
0.185
0.19
Population
In Range
2402042
1865764
1182914
669054
501790
317283
224167
167263
150020
139674
72423
67250
72423
36212
65526
17244
44834
31039
18968
20692
22417
20692
12071
10346
10346
13795
13795
15519
8622
8622
5173
3449
1724
6897
3449
8622
6897
1724
Lower
Risk
Estimate
20.80
32.31
30.72
23.17
21.72
16.48
13.58
11.58
11.69
12.09
6.90
6.99
8.15
4.39
8.51
2.39
6.60
4.84
3.12
3.58
4.08
3.94
2.40
2.15
2.24
3.11
3.22
3.76
2.16
2.24
1.39
0.96
0.49
2.03
1.04
2.69
2.21
0.57
Upper
Risk
Estimate
83.18
129.22
122.89
92.68
86.88
65.92
54.34
46.34
46.76
48.37
27.59
27.95
32.60
17.56
34.04
9.55
26.39
19.35
12.48
14.33
16.30
15.76
9.61
8.60
8.96
12.42
12.90
15.05
8.66
8.96
5.55
3.82
1.97
8.12
4.18
10.75
8.84
2.27
Overall Risk Estimate
* Lung Cancer Deaths/Year
Lower Upper
376 1505
t
(continued)
23
-------
Approximate Yearly Lung Cancer Deaths from
Exposure to Airborne Radon In New England (continued)
pCI/L Fraction
Ranae In Range
39 0.000825
40 0.000618
41 0.000412
42 0.000618
43 0.000412
44 0.000412
45 0.000618
46 0.000206
48 0.000412
50 0.000618
52 0.001031
53 0.000412
54 0.000206
55 0.000206
56 0.000206
58 0.000206
59 0.000412
60 0.000206
61 0.000206
62 0.000412
64 0.000206
66 0.000206
72 0.000412
76 0.000206
81 0.000206
84 0.000206
87 0.000206
90 0.000206
98 0.000206
100 0.000206
106 0.000206
109 0.000206
115 0.000206
116 0.000206
122 0.000206
124 0.000206
130 0.000206
135 0.000206
158 0.000206
187 0.000206
207 0.000206
243 0.000206
326 0.000206
335 0.000206
417 0.000206
WL
Range
0.195
0.2
0.205
0.21
0.215
0.22
0.225
0.23
0.24
0.25
0.26
0.265
0.27
0.275
0.28
0.29
0.295
0.3
0.305
0.31
0.32
0.33
0.36
0.38
0.405
0.42
0.435
0.45
0.49
0.5
0.53
0.545
0.575
0.58
0.61
0.62
0.65
0.675
0.79
0.935
1.035
1.215
1.63
1.675
2.085
Population
In Range
6897
5173
3449
5173
3449
3449
5173
1724
3449
5173
8622
3449
1724
1724
1724
1724
3449
1724
1724
3449
1724
1724
3449
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
1724
Lower
Risk
Estimate
2.33
1.79
1.22
1.88
1.28
1.31
2.02
0.69
1.43
2.24
3.88
1.58
0.81
0.82
0.84
0.87
1.76
0.90
0.91
1.85
0.96
0.99
2.15
1.13
1.21
1.25
1.30
1.34
1.46
1.49
1.58
1.63
1.72
1.73
1.82
1.85
1.94
2.02
2.36
2.79
3.09
3.63
4.87
5.00
6.23
Upper
Risk
Estimate
9.32
7.17
4.90
7.52
5.14
5.25
8.06
2.75
5.73
8.96
15.53
6.33
3.22
3.28
3.34
3.46
7.05
3.58
3.64
7.40
3.82
3.94
8.60
4.54
4.84
5.02
5.20
5.37
5.85
5.97
6.33
6.51
6.87
6.93
7.29
7.40
7.76
8.06
9.43
11.17
12.36
14.51
19.47
20.00
24.90
Overall Risk Estimate
# Lung Cancer Deaths/Year
Lower Upper
(continued)
-------
Approximate Yearly Lung Cancer Deaths from
Exposure to Airborne Radon In New England (continued)
EPA equation: Lung Cancer Deaths = CR * T * FWLM * RRRM * TCR * POP
WL/pCi/L= 0.005 Conversion from pCi/L to Working Level
(assumes 50% equilibrium).
Population:
Single = 6847000
Duplex = 1511000 Note:Actual 2-4 family units = 3022000.
Use 1/2 to estimate side-by-side duplexes.
Const = 0.57716 = T * FWLM * TCR
RRRMlow = 0.01
RRRMhi = 0.04
Liv/Bas = 0.3 Assumes the lived in area is, on average
only 30 X of the basement concentration.
Note: Uses the calculation in the EPA Radon Reference manual.
Uses data on radon measurements in basements for the six New England states and
EPA/Region I.
25
-------
5. Indoor Air Pollutants Other than Radon
Problem Area Definition
This category includes exposures to air pollutants in schools, offices, homes, and indoor
spaces from sources in the indoor environment. Common sources of indoor air pollution
include cleaning supplies, paints, gas stoves, pressed wood products, upholstery, pesticides,
and cigarettes. Indoor air pollutants include environmental tobacco smoke, formaldehyde,
benzene, carbon monoxide, carbon tetrachloride, trichloroethylene, oxides of nitrogen,
chlordane, heptachlor, and phthalate esters.
Summary/Abstract
A draft EPA report, Indoor Air Pollution, The Magnitude and Anatomy of Problems and
Solutions: A Scoping Study (1987), served as the resource for much of the exposure
information. Information presented in this report, which summarized monitoring information
and reporting time weighted exposure levels by building type, was relied on quite heavily.
This information was combined with Region I specific population statistics, obtained from the
census bureau, in order to evaluate both carcinogenic and non-carcinogenic (acute and
chronic) health effects. The computed upper-bound estimate of the individual lifetime cancer
risk from exposure to indoor air pollutants was approximately 1E-2 and the population risk
for Region I was approximately 1E-3 (annual incidence of 50). The chronic non-cancer health
effects were projected for a large population but were of relatively low potency (exposure
levels were only slightly greater than the standard).
The computed cancer and non-cancer health effects assume simultaneous exposure to each
of the nine carcinogens and three non-carcinogens evaluated, which is most likely a
conservative assumption. Additionally, the number of individuals in New England who
actually have sources of pollution in their indoor environments and those at risk of
experiencing adverse health effects were extremely difficult to estimate and may not be an
accurate representation. The cancer potency factors and the health effects associated with the
non-carcinogens have been studied quite extensively. In general, we are confident about the
hazard identification portion but less certain about the exposure assessment The percentage
of problem covered was high due to the relatively large number of chemicals evaluated.
26
-------
Toxicity Assessment
Selecting just a handful of chemicals for evaluation from the many hundreds of indoor air
pollutants we are exposed to every day was a somewhat difficult task. We attempted to select
chemicals on the basis of their toxic effects, the availability of exposure information, their
prevalence in indoor environments, and their ability to serve as surrogates of exposure for
other, similar chemicals.
The known or suspected human carcinogens selected for evaluation include the following
volatile organic chemicals: benzene, carbon tetrachloride, tetrachloroethylene,
trichloroethylene, and formaldehyde. Pesticides evaluated include chlordane and heptachlor.
Environmental tobacco smoke and phthalate esters were also evaluated for carcinogenic
effects.
Substances evaluated for their non-carcinogenic effects include carbon monoxide, oxides
of nitrogen, respirable participates, and formaldehyde.
Health Effects of Concern
Non-Carcinogenic
The National Ambient Air Quality Standards (NAAQS) for CO, NOX, and respirable
particulates served as the standard for acute and chronic exposures. A recommendation put
forth by the American Society of Heating, Refrigerating, and Air Conditioning Engineers
(ASHRAE) for formaldehyde was used to assess the potential for health effects due to this
substance.
The non-carcinogenic endpoints that formed the basis for our standard-setting process
(NAAQS) were selected as the health effects of concern for this problem area. The health
effects of concern associated with chronic and acute exposures to indoor air pollutants include
cardiac impairment (CO); increased respiratory illness and decreased lung function (NO»);
impaired lung defense mechanisms, increased lung resistance, and a worsening of pre-existing
lung disease (respirable particulates); and eye and respiratory irritation (formaldehyde).
Health effects that may occur at levels of exposure below the acute/chronic standards were not
evaluated.
Cancer Potency
Preference was given to using potency factors for the inhalation route of exposure, but in
the absence of a potency factor based on an inhalation study, oral potency factors were used.
The potency factor for environmental tobacco smoke was not assessed. Instead, cancer risk
estimates were based on incidence data generated by leading researchers in this field. Potency
factors expressed in (mg/kg/day)-l are as follows:
27
-------
benzene 0.029
carbon tetrachloride 0.13
tetrachloroethylene 0.0017
trichloroethylene 0.0046
formaldehyde11' 0.038
phthalate esters** 0.00067
chlordane 1.3
heptachlor 4.5
* Based on a value reported in the EPA 1987 Assessment of Health Risks to Garment
Workers and Home Residents from Formaldehyde.
** Based on unit risk for di-(ethylhexyl)-phthalate.
Exposure Assessment
Common sources of indoor air pollutants include cleaning supplies (solvents), paints, gas
stoves, combustion sources, pressed wood products, upholstery, urea formaldehyde foam
insulation, pesticides, and cigarettes. These sources are in our homes, offices, schools, and
public buildings.
For this category, we considered direct inhalation exposures resulting primarily from the
volatilization or off-gassing of pollutants generated from activities that take place in indoor
environments (i.e., burning wood or a cigarette produces particulates, formaldehyde, NOX,
CO. and benzene).
As mentioned above, a draft EPA report on indoor air pollution served as the basis for
much of the exposure information. Assumptions used in the exposure assessment were
heavily based on the assumptions from the draft EPA report As a result, three common
indoor exposure environments were selected as the foci for this problem area: homes, offices,
and schools. Because people readily move about, and the pollutants and the exposure
concentrations differ from location to location, we took into consideration human activity
patterns when estimating average daily exposures. In order to do this, we assumed that we all
breathe, on average, 1.3 nr of air each hour and that over the course of a lifetime we spend:
-16 hrs/day at home every day
- 6 hrs/day at school, 180 days/yr for 12.6 years
- 7.5 hrs/day at the office, 250 days/yr for 45 years
Exposure point concentration ranges were selected based on typical levels reported in the
literature, many of which were summarized in the draft EPA report on indoor air pollution.
They were selected to reflect the average levels one might find if a particular source were
present We have tried to exclude pollutant levels representative of worst-case situations from
our analysis.
28
-------
Because each pollutant evaluated can be associated with a particular source and because
individuals without exposure to a pollutant source are likely to have a much lower risk, the
population at risk depended on the pollutant and the number of individuals exposed to that
particular source or demonstrating a certain behavior. Additionally, for the non-cancer health
effects, the population at risk was often a subset of those individuals potentially exposed to
any given source, usually those who are more susceptible to the health effects of concern.
U.S. census information served as a basis for much of the population data. The estimated
population for the six New England states is approximately 12 million. The estimated figure
representing the civilian labor force is 6 million, and approximately 2 million people in New
England are in school grades K through 12.
Risk Characterization
The compounds associated with the greatest individual risk for cancer were environmental
tobacco smoke, chlordane, phthalate esters, formaldehyde, and heptachlor. These compounds
were also associated with the greatest population risk. Exposures in the home environment
contributed most significantly to the overall cancer risk estimate. The total estimated
individual cancer risk from exposure to all pollutants evaluated ranged from 0 to 1E-2, and the
total population cancer risk estimate for all nine pollutants evaluated ranged from 0 to 1E-3.
The corresponding annual incidence is between 0 and SO.
For the evaluation of chronic non-cancer effects, estimated exposures to CO, NOg, and
CHoO only slightly exceeded the standard and thus each received a low ratio score. The
health effects of concern at the respective standards varied in severity from mild to
moderately severe. In general, large populations were deemed to be potentially exposed to the
pollutant levels evaluated.
Acute non-cancer effects were also evaluated for CO, NOX, and respirable particulates.
The severity and potency scores remained unchanged from the chronic evaluation but a
smaller population was considered at risk. The acute effects did not warrant a change in our
thinking on the importance of the non-cancer health effects.
' One source of uncertainty is the conservative approach taken in assuming that people may
actually be exposed to every pollutant evaluated at the levels reported. Anyone we could
associate with a certain behavior or source was assumed to be exposed to the average levels
reported. Estimates of sensitive populations and populations at risk (demonstrating a certain
behavior or having a particular source) proved to be very difficult Consequently, much of the
population information was collected from census statistics. Because we had very little
monitoring information from the Regional Office, levels of exposure were based on typical
concentrations found in indoor environments from studies done across the nation. While not
specific to Region I, we felt the levels probably do not change radically from one location to
another. Non-carcinogenic health* effects at levels below standards were not considered in this
evaluation.
29
-------
We felt the percentage of problem covered was high because of the relatively large
number of pollutants covered. We did not evaluate every indoor air pollutant for the problem
area but we tried to select those pollutants that represented different sources and chemical
groups, and ones for which we had exposure information.
30
-------
6. Radiation from Sources Other than Radon
Problem Area Definition
This problem area includes risks from exposure to radiation other than indoor radon.
Various sources of ionizing radiation (such as occupational exposure, nuclear power plants,
and low level radioactive wastes) were eliminated from this problem definition because of
the lack of regional EPA involvement in managing them. The problem area was then
expanded to include sources of non-ionizing radiation. Although regional involvement in
these sources might also be limited, it was thought that the potential for generally
unrecognized risks should be explored through the Risk Reduction Project.
Two general sources of non-ionizing radiation are of concern in non-occupational
settings: radio frequency radiation (mainly FM radiation in the vicinity of FM antennas),
and extremely low frequency (ELF) radiation-die electromagnetic radiation that is found in
the vicinity of 60 Hz power sources. ELF has recently been the subject of both laboratory
research and epidemiological studies. Because of the pervasiveness of ELF and the
availability of some quantitative estimates of risk from ELF, it was decided to analyze this
problem in terms of risks from ELF radiation. It should be recognized, however, that there
are other sources of non-ionizing radiation and ionizing radiation that are not considered in
this evaluation. Regional data were not evaluated for this problem area. Rather, this paper
discusses the implications of current research on the risk posed by ELF radiation.
Summary/Abstract
The most comprehensive review of the effects of powerlines is the New York State
Power Lines Project (July, 1987). This $5 million project reviewed existing research and
funded additional studies to evaluate a large variety of possible health effects that could be
caused by electric and magnetic fields resulting from transmission lines. Studies included an
evaluation of teratogenic effects, effects on cell biology, effects on the proliferation of cancer
cells in culture, neurobiological effects, behavioral effects, multidisciplinary studies of
human exposure, and the epidemiology of cancer incidence.
Several areas of possible concern were identified, but a study by Savitz that
demonstrated a possible association between residential magnetic fields and the incidence of
childhood leukemia in Denver was of particular concern. In the Savitz case-control study,
all cases of childhood cancer (ages 3-14) between 1978 and 1983 were evaluated and
compared with a control group. Exposures to magnetic fields were estimated by determining
wiring configurations and by direct measurement of fields. Results showed that there was an
association between wiring configuration and increased cancer risk. This held for all
childhood cancers, especially for leukemias, and, to a lesser degree, for brain tumors. The
relative risk was 2 for the highest exposed group.
31
-------
According to the study, the incidence rate for childhood cancers is about 1 per
10,000 children per year. If the relationship shown by Savitz is true, then the risk for the
most exposed individuals would be increased to 2 per 10,000 children per year. The study
concludes that this could mean that 10 to IS percent of childhood cancers are attributable to
magnetic fields. An EPA estimate of the number of cases nationally (using the IS percent
risk estimate) is 300 cases per year attributable to ELF.
Although the Savitz study is the most conclusive epidemiological study conducted, there
have been others as well. These studies were criticized because they measured only
powerline proximity and not the actual fields in the homes. More definitive analytical
epidemiological studies are necessary to strengthen this potential association.
Toxicity Assessment
The "pollutant" in this case is the magnetic field that is generated by 60 Hz powerlines.
The mechanism of causation is not known, but it is possible that these fields may promote
cancer by stimulating growth-related hormones in cancerous cells. The cell-stimulating
effects of electromagnetic fields have been studied for years in efforts to increase cell activity
to heal bone fractures. In a separate powerlines project study, researchers reported that
cancer cells reproduced faster after exposure to electromagnetic fields.
Exposure Assessment
Fields generated by 60 Hz powerlines are ubiquitous, and are generated by any type of
powerline, from large transmission lines to smaller distribution lines to household wiring.
Magnetic fields are measured in units of the tesla or the gauss. In the Savitz study, the risk
of leukemia increased with measured magnetic fields, with the relative risk of getting
leukemia at 2.1 for fields greater than or equal to 0.2S microtesla (2.S milligauss).
An extrapolation from the Savitz study to the rest of the country would be based on the
assumption mat the exposure patterns found in the Denver area would be the same
everywhere else. The population studied by Savitz was not large-only 3S7 cases.
The population at risk is defined to include children between the ages of 3 and 14.
32
-------
Risk Characterization
The risk estimates derived from the Savitz study indicate that the relative risk of
childhood cancer is twice as great for children in high-exposure homes. The risk increases
from 1 in 10,000 to 2 in 10,000.
The uncertainty about these risk estimates is high. However, the estimates do form the
basis for a hypothesis that childhood cancer is promoted by exposure to ELF fields.
The overall estimate on the percentage of the problem covered is medium.
Issues not considered include the following:
• Effects of ELF on adult populations
• Effects of RF fields
• Effects of ionizing radiation from sources not considered in the problem definition
33
-------
7.8. & 9. Industrial Point Source, POTW,
and Nonpoint Source
Discharges to Surface Waters
Problem Area Definition
The three problem areas dealing with discharges to surface water were combined for this
analysis. Because contamination may be the result of discharges from several different
sources, it is difficult to clearly separate these categories and attribute problems to a particular
source.
The sources considered in this analysis include point source discharges from direct
industrial dischargers; discharges from POTWs that may include wastewater from municipal
sewage treatment systems, wastewater from industrial indirect dischargers, and combined
sewer overflows; and nonpoint discharges that do not originate from a discrete conveyance or
pipe but may be due to runoff, discharge from contaminated ground water, and suspension or
dissolution of contaminated sediments.
Summary/Abstract
Using data on fish contamination in areas where fish advisories are in effect, we calculated
a cancer incidence rate of 16 to 32 cases annually due to PCB contamination. The individual
exposure risk is approximately 7.6E-3 to l.SE-2. The population exposed is 150,000, an
estimate of the number of recreational fishermen in the region. For non-cancer risks, the
estimated doses for lead and mercury are 10 times and 25 times the oral RfD, respectively.
This relatively high cancer and non-cancer risk is attributed predominantly to discharges from
industrial and nonpoint sources. From the information available, POTWs did not contribute
significantly to the contamination considered and therefore received a lower ranking.
Toxicity Assessment
The major chemicals of concern in this analysis include lead, mercury, PCBs, and dioxin.
These are the substances on which fish consumption advisories have been based in
Massachusetts and Connecticut. The RfDs (mg/kg/day) for the non-carcinogens are 6E-4 for
lead and 5.1E-5 for chronic effects from mercury. The health endpoints associated with lead
and mercury include increased blood pressure, reproductive effects, and central nervous
system effects. The cancer potency factors used in the analysis include 7.7 per mg/kg/day for
PCBs and 1.56E-5 per mg/kg/day for dioxin.
34
-------
Exposure Assessment
As discussed previously, there are numerous potential sources of contamination; the
specific type, size, and location of sources may not be readily identified. Our information
does, however, allow us to draw some conclusions about the relative importance of industrial,
POTW, and nonpoint sources.
The primary pathway of exposure considered in this analysis is ingestion of contaminated
fish from surface waters. The health risks associated with inhalation and direct contact with
surface water are assumed to be minimal. A real risk may be associated with the direct
ingestion of surface water. However, this risk is difficult to assess. Because the population
consuming untreated surface water is small and water concentrations of contaminants are low,
the risk is assumed to be significantly less than that associated with fish ingestion.
Several assumptions were used in characterizing the exposure to contaminants. We
assumed that there is no risk from areas in which fishing has been legally restricted. We only
considered areas where people might continue to fish even though fish consumption
advisories are in effect Exposure calculations were based on average concentrations of
contaminants in fish from data in Massachusetts and Connecticut In cases where average
concentrations across sampling stations or species varied significantly, we used a range of the
average measured concentration.
The exposure assessment also uses a typical consumption figure of 46 grams/day that is
based on the average consumption offish among Caucasian marine sport fishermen and
women in New England, a statistic from the National Marine Fisheries Service. We assumed
that the consumption by non-marine recreational fishermen and women would be roughly the
same.
We calculated that approximately l.S million people in Region I fish recreationally. This
number is based on data from the National Marine Fisheries Service and on a survey. Of this
population, however, we assume that only 10 percent of the recreational fishermen and
women would be fishing in areas where contamination may be a problem, i.e., areas where
fish consumption advisories are in effect or could potentially be issued.
The average concentration values and estimated consumption value allowed us to
calculate typical doses (mg/kg/day) of contaminants:
Cone, in Average Average
Dose = fish tissue x consumption divided body weight
(mg/kg/day) (mg/kg) (kgs/day) by (70kgs)
35
-------
Risk Characterization
Using these exposure assumptions, we calculated cancer and non-cancer risk estimates.
The cancer risk estimate associated with PCBs is 7.6E-3 to 1.5E-2 (using a range of the
average concentrations of PCBs in fish). The cancer risk from dioxin is minimal (1E-9). The
individual cancer risk for PCBs translates into a population risk of 1,140 to 2,250 cancer cases
over a lifetime. Using a 70-year average lifespan, die annual cancer incidence associated with
PCBs is 16 to 32 cases.
The non-cancer risk from ingestion of contaminated fish is due to consumption of lead
and mercury. The typical dose of lead is calculated to be 5.8E-3 mg/kg/day. This is
approximately 10 times the RfD for lead. For mercury, the typical dose ranges from 3.3E-4 to
1.3E-3 mg/kg/day. The upper range of this dose is approximately 25 times the RfD for
chronic effects. The health endpoints associated with lead and mercury include increased
blood pressure, reproductive effects, and central nervous system effects. These are assigned a
severity score of 3. The population score corresponding to 150,000 people is 3. Finally, the
potency score associated with doses ranging from 10 to 25 times the RfD is 2. The total score
for non-cancer therefore is 8.
These risks are assumed to be associated predominantly with contamination from
industrial discharges and nonpoint sources. The contribution from POTWs is assumed to be
minimal, thus the score for POTWs is lower than that for industrial and nonpoint sources.
Based on this analysis, the uncertainty in the risk estimates is considered to be medium.
The percentage of the problem covered, however, is assessed as low because only a small
number of pollutants was evaluated.
36
-------
10. Discharges to Estuaries, Coastal Waters,
and Oceans from All Sources
Problem Area Definition
This problem area assesses risks from exposure to contaminants discharged to estuaries,
coastal waters, and oceans from all sources. These sources may include runoff, ocean sludge
disposal, combined sewer overflows, POTW wastewaters, air deposition, and others.
Summary/Abstract
Using data on contaminant concentrations in shellfish and finfish measured throughout
New England, we calculated a total individual cancer risk of 4.9E-4. This cancer risk
represents the sum of individual risks from nine contaminants. Risks from individual
chemicals range from approximately 2E-4 for heptachlor and PCBs to 5E-8 for DDT and
hexachlorobenzene (HCB). Assuming the total New England population of 12.8 million
people is exposed, this risk would translate into approximately 6,272 cancer cases over a
lifetime (70 years). The annual incidence is therefore 90 cases.
Non-cancer risks are not expected. None of the 18 pollutants assessed exceeded their
respective reference doses (RfDs).
Toxicity Assessment
The data we examined included 18 different pollutants: PCBs, silver, cadmium,
chromium, copper, iron, mercury, nickel, lead, zinc, DDT, dieldrin, heptachlor, lindane, PAH,
chlordane, hexachlorobenzene, and hexachlorocyclohexane. The cancer potency factors for
these pollutants are listed in Table 10-1, along with the RfDs, typical concentrations, and
calculated doses. The doses calculated are all significantly below the oral RfDs, with the
exception of lead. The exposure dose for lead approached one-half the RfD. The various
non-cancer health endpoints associated with the RfDs were therefore not addressed.
Exposure Assessment
There are numerous potential sources of contamination; the specific type, size, and
location of sources cannot be readily identified. The levels of contaminants found in shellfish
and bottom fish generally correlate well with sediment concentrations. Sources of sediment
contamination are not easily identified but often result from industrial point sources, CSOs, or
wastewater discharges. This analysis did not attempt to attribute contamination to particular
sources.
37
-------
Table 10-1
Estimated Risk Associated with F/sA Contamination
RGBs
Silver
Gadnrium
Chromium
Copper
Iron
Mercury
Nickel
Lead
Zinc
DDT
Dieldum
Hoptschlor
Undane(g-HCH)
PAH
Chlordane
HCB
HCH
The onlv route c
Concentration
In Hah
(mp/kg)
0.1
0.03
0.2
0.4
3.0
35.0
0.03
0.3
1.0
25.0
0.006
0.02
0.2
0.002
0.01
0.003
0.00013
0.00015
)fexoosurec<
Typical
Doae C.P.F. „ Cancer
(ma/Kg-day) (ing/kg-day)'1 Risk
2.3E-5 7.7 1.8E-4
7.0E-4
4.6E-5
9.1E-5
6.9E-4
8.0E-3
7.0E-6
6.9E-5
2.3E-4
5.7E-3
1.4E-7 0.34 5.0E-8
4.6E-6 16.0 7.3E-5
4.6E-5 4.5 2.1E-4
4.6E-7 1.3 5.9E-7
2.3E-6 11.5 2.6E-S
6.9E-7 1.3 8.9E-7
3.0E-8 1.67 5.0E-8
3.0E-8 6.3 2.2E-7
jnsidered in this evaluation was incestion of o
RID
(mg/kg-day)
1.0E-4
3.0E-3
2.7E-4
5.0E-3
3.7E-2
2.0E-3
1.4E-3
2.1E-3
5.0E-4
3.0E-4
5.0E-5
6.0E-4
3.0E-4
ontaminated
fish and shellfish. Some risk may also be associated with direct contact or swimming in
contaminated water, however, this is considered minimal in comparison with risks from
ingestion.
Several assumptions were used in characterizing the exposure to contaminants. We used
"typical," or median, concentrations of pollutants, based on data from various studies
summarized in a report by the Woods Hole Oceanographic Institute and the Environmental
Science Program at the University of Massachusetts. This report contains measurements of
fish tissue and shellfish concentrations for a number of contaminants, species, and sampling
locations throughout New England. We have also used contaminant levels from a separate
report on Quincy Bay. In Quincy Bay, the typical fish concentrations listed for cadmium,
38
-------
chromium, copper, mercury, lead, and DDT were all slightly less than the concentrations we
chose as "typical" from our other New England data. The mean levels of PCBs and PAHs,
however, were higher than the median concentrations listed in the New England data. For all
concentrations, therefore, we relied on typical concentrations from the New England data,
except in the case of chlordane, hexachlorobenzene, and hexachlorocyclohexane, for which
we only had concentration data from Quincy Bay.
It is important to keep in mind that when we determined typical or median concentrations,
we excluded data taken from New Bedford Harbor, since all fishing and shellfishing in the
harbor has been banned. In this area, concentration observed may not be considered as
representative of the exposure concentration. This may or may not be an accurate assumption.
It is assumed in this analysis that other measurements of concentrations from the New
England waters represent the typical concentrations observed in Region I.
The typical exposure dose of the contaminants was calculated by using a typical
consumption rate of 16 grams/day for the New England population. This rate represents an
average consumption for all the people in New England, as determined by the National
Marine Fisheries Service. The typical exposure doses calculated for an adult (70 kg body
weight) are shown in Table 10-1. The population exposed is assumed to be the total New
England population, or 12.8 million people.
Risk Characterization
Cancer risk estimates associated with exposure to several pollutants are presented in
Table 10-1. Assuming that cancer risks should be additive, the total individual cancer risk is
calculated as 4.9E-4. This translates into a total of 6,272 cancer cases over a lifetime (70
years) for the entire New England population, or an annual incidence of 90 cases. Non-cancer
risk is not expected to occur.
This estimation of annual cancer cases probably represents an upper bound estimate.
Calculation of the median concentrations may involve some hot spots of contamination. The
uncertainty associated with these estimates is high. The percentage of the problem covered,
however, is medium, because of the number of contaminants considered.
39
-------
12. Drinking Water
Problem Area Definition
As drinking water arrives at the tap, it may contain a wide variety of contaminants from
both natural and man-made sources. This category covers water supplies from both surface
water and ground-water sources. It relates to watershed management and drinking-water
contamination from the source to the tap. Exposure may be through ingestion, inhalation of
volatilized contaminants, and dermal contact It excludes contamination from waste sites
(problem areas # 13, # 14, # IS, # 16) and storage tanks (#18). The impact of lead in drinking
water is evaluated in the Lead problem area (#22) and that of radon from drinking water is
evaluated in the Radon problem area (#4).
Summary/Abstract
Individual cancer risk derived from exposure to drinking water contaminants evaluated in
this analysis is approximately 2E-4. The annual excess cancer cases attributable to
drinking-water contaminants are estimated to be lower than 20. Doses rarely exceed the
Reference Dose (RfD) or Maximum Contaminant Level (MCL). When they do, it is on the
order of one to five times the benchmark. This analysis is based on monitoring data, data on
violations of MCLs from the Federal Reporting Data System (FRDS), and information from
the Federal Register (Advanced Notice of Proposed Rulemaking for Radionuclides).
The quality of the data used in this analysis is considered inadequate. The percentage of
the problem covered is medium and the uncertainty is high.
Toxicity Assessment
The primary chemicals of concern identified for this problem area in Region I are some of
those contaminants regulated or considered by EPA for regulation. These include arsenic,
copper, microbial contamination, total trihalomethanes (TTHMs), nitrates, andradionuclides
(excluding radon).
. 40
-------
Major sources of pollutants evaluated are:
• Natural sources for arsenic and radionuclides.
• Corrosion of potable pipings and joints by corrosive water for copper and lead.
Although lead is also a major contaminant of concern in Region I, it is evaluated
in the Lead problem area (#22).
• Point sources from effluent discharges, sewer overflow, sewage treatment plants,
and failure of septic systems and nonpoint sources for microbial contaminants.
• Application of lawn-care or agricultural fertilizers and failure of septic systems
for nitrates.
• Disinfectant byproducts from chlorination for TTHMs.
Of the contaminants evaluated, arsenic and radionuclides are human carcinogens
(Class A). Chloroform is the predominant compound (>80 percent) among the TTHMs
derived from chlorination of drinking water. Chloroform is considered a probable
human carcinogen (Class B2). The Cancer Potency Factor (CPF, or slope factor) used
for this analysis (see Table 12-1) is assumed to be the same as that derived for
chloroform (8.33E-6 per mg/L). The CPF for arsenic is 4.29E-S per mg/L, and the CPF
for radium-228 is 8.8E-6 per pCi/L.
For non-carcinogenic effects, the reference dose (RfD) is used in this evaluation.
Where the RfD is not applicable or available for the contaminants, the relevant MCL is
used. The RfDs for copper, chloroform, and nitrates are 0.037 mg/kg/day,
1 mg/kg/day, and 0.01 mg/kg/day, respectively.
Other contaminants were evaluated against their MCLs as promulgated by EPA.
The MCL for arsenic is 0.05 mg/L, and the MCL for coliform is 1 count per 100 ml
sample analyzed. There is no MCL for Giardia. The toxicity endpoints associated with
the contaminants are shown in Table 12-1 along with their respective CPF, RfDs and
MCLs.
41
-------
Table 12-1
Toxlcologlcal Parameters of Drinking Water Contaminants
Contaminants
Arsenic
Copper
Nitrates
TTHMs*
Radium-228
Coliform
Giardia
Carclnogenlclty
Classification
A
-
-
B2
A
-
_
Cancer Potency
Factor
4.29E-5
(ug/L)-l
NA
NA
8.33E-6
(ug/L)-1
8.8E-6
NA
NA
Reference
Dose
-
0.037
mg/kg/day
1.0
mg/kg/day
0.01
mg/kg/day
-
-
<1 cyst
MCL
0.05
mg/L
-
10
mg/L
100
ug/L
5pCi/L
1 count
per 100 ml.
Toxlcologlcal
End Points
Cancer
Gl irritation
Methemoglobinemia
Cancer
Cancer
Gl Irritation
Gl Irritation
* Use chloroform as the surrogate compound. CPF and Rf D derived from ingestlon of chloroform
administered in drinking water is used.
- Not available.
NA: Not applicable.
42
-------
Exposure Assessment
The data sources for evaluating exposure to drinking-water contaminants varied.
Monitoring data are available from state agencies or their laboratories, but due to time
constraints and resource limitations this evaluation only used data that were easily accessible.
Information from the FRDS for MCL violations during 1987 were examined. The only
violating parameter was coliform. Although there were rarely MCL violations for TTHMs,
Region I nonetheless is concerned with measured concentrations below MCL of 100 ug/L.
For this analysis, we examined TTHM monitoring data compiled by Region I's Water Supply
and Ground Management Branch from data submitted by all six New England states. A
random sample of SO systems was selected to compute median values for each state and for
New England as a whole. For arsenic, copper, and nitrates, we examined sampling results
from a state laboratory in New Hampshire. Radium data were based on the regional
information published in the Federal Register, Vol. 51, No. 189,34836-34862, September 30,
1986. The Giardia information was based on an interview with staff from the EPA
Laboratory in Lexington, Massachusetts, and a selection of newspaper and journal articles.
The median concentrations of contaminants, which are considered to be representative of
typical conditions, are presented in Table 12-2.
Table 12-2
Estimated Median Concentration of Contaminants
Found In Drinking Water
Contaminants Median Concentration
Arsenic <5 ug/L
Copper 100 ug/L
Nitrates 0.25 mg/L
TTHMs 49 ug/L
Radium-228 0.10pCi/L
Coliform 1 -5 counts pen 00 ml
Giardia Positive testing
Outbreaks of Giardia were documented in at least four surface-water supply systems in
New England. It was estimated that 10 percent of all finished water sampled from
surface-water supplies tested positive for Giardia. It should be noted that those tests where
Giardia was not isolated from the sample do not guarantee that the water supply was negative
for Giardia. It simply means the analyst was unable to isolate Giardia from the sample.
Giardia cysts are considered to be a threat to virtually all water utilities using surface waters.
43
-------
Routes of exposure to contaminants in drinking water include ingestion, inhalation, and
dermal absorption. However, the only route of exposure quantitatively evaluated in this
analysis is ingestion. Inhalation of TTHMs released from all uses of contaminated water in
the household could possibly contribute as much exposure dosage as that due to ingestion.
Therefore, the cancer risk estimate was doubled from what was calculated for the ingestion
route. Dermal absorption can also contribute additional exposure dose. However, it is
believed to be not greater than 25 percent of that contributed by the ingestion route in the
worst case. Dermal exposure is not quantitatively evaluated.
The exposed population varied depending upon the contaminants examined (Table 12-3).
For TTHMs, the exposed population was estimated to be 7 million (those using surface-water
supplies). For arsenic and radium, the population was determined to be 6 million
ground-water users. For contaminants where violations were evaluated (i.e., colifonn, copper,
and nitrates), the exposed population was estimated as only those serviced by water systems
with violations. For colifonn and copper, there were approximately 40 violating systems
across the Region I; the population exposed was determined to fall into the 10,000 to 100,000
range. Similarly, the population was considered to fall in the 1,000 to 10,000 range for
children exposed to public water systems violating the nitrates MCL.
A study conducted by the Colorado State University (1986) examined the presence of
Giardia in raw and finished surface water and ground water from 300 sites in 28 states.
Approximately 10 percent of all finished surface water supplies tested positive for Giardia.
Combining this information with the total number of surface water users in Region I
(7 million) led to our estimate of 700,000 people exposed to Giardia.
Risk Characterization
In the evaluation of daily exposure doses, a daily ingestion of two liters of water for 70-kg
adults for the lifetime of 70 years is generally assumed. For nitrates, one liter of daily water
consumption is assumed for 10-kg children.
Chronic daily intakes (GDI) are calculated as follows:
CDI= [Contaminant concentration] x 2 L
70kg
or in the case of children exposed to nitrates:
CDI = [Contaminant concentration] x 1L
10kg
44
-------
Table 12-3
Characteristics and Size of Population
Exposed to Drinking Water Contaminants
Population . Number of People
Contaminants Characteristics Exposed
Arsenic Ground-water users 6,000,000
Copper Users of public 10,000-100,000
water supply violators
Nitrates Children served by 1,000-10,000
public water supply
violators
Radionudides Ground-water users 6,000,000
TTHMs Surface water users 7,000,000
Coliform Users of public 10,000 -100,000
water supply violators
Glardla Users of surface 700,000
water tested positive
For non-carcinogenic effects, The Hazard Index (HI) is a comparison of GDI to the
relevant RfD considered protective of adverse health effects.
Hazard Index (HP = GDI
Reference Dose*
* MCL is used for colifonn and arsenic.
Assuming additivity, the HI for contaminants with similar modes of action for toxicity or
similar toxicity endpoints can be summed.
Incremental lifetime individual cancer risk is a function of GDI and chemical specific
CPF.
Individual cancer risk = GDI x CPF
Annual cancer incidents Individual cancer risk x number of
in Region I = people exposed to the contaminants
in the region/70 years
Table 12-4 presents the estimated HI, the individual lifetime cancer risk, and the annual
cancer incidence in New England resulting from exposure to drinking contaminants evaluated
in this analysis.
45
-------
Individual cancer risk was estimated at 1.6E-5 for surface water users, based on the
median TTHM concentration (49 ug/L) across the region. Risk from exposure to arsenic was
estimated at the detection limit (a limitation of the monitoring data used) for an individual risk
of 2.1E-4. Radium risks were estimated using average values for the six Region I states at
8.8E-6.
Table 12-4
Risk Characterization of Drinking-water Contaminants Examined
Lifetime Individual Annual Cancer or
Contaminants HI Cancer Risk Disease Incidence**
Arsenic <1
Copper <1
Nitrates <1
TTHMs <1
Radium-228 t
Coliform 5
Giardi
2.1 E-4
NA
NA
1.7E-5
8.8E-7
NA
1E-2*
<17
t
t
2
«i
t
. 7,000
•Individual risk for Giardiasis.
** Annual incidence estimated In Region I.
t - Not analyzed
NA-Not applicable.
46
-------
The population risks were estimated using the exposed populations discussed above. For
TTHMs, median concentration values and surface-water-dependent populations specific to
each state were used to develop the population risk. The resulting estimate was about
120 cancer cases in 70 years, or approximately two cases per year. Arsenic (at the detection
limit) results in an estimate of approximately 17 excess cancer cases per year. Radium
exposure results in an estimated less than one cancer case per year. The estimate of total
cancer cases is less than 20 per year.
For the non-carcinogenic effects, the estimated GDI are below their respective RfDs.
Since these chemicals do not exhibit similar systemic toxic effects, His from individual
contaminants were not considered additive. The coliform violations in Region I had a median
value of five times the MCL. Incidence data from two recent outbreaks of Giardiasis were
used to characterize the risks in that area. Both case studies reported an individual disease risk
of 1E-2.
The percentage of the problem covered is considered medium and the uncertainty is high.
Although this analysis primarily relied on monitoring data and captured most of the
contaminants of concern in this region, the accessibility of the data is limited. Also, the
quality of the data is not uniform across the contaminants examined.
While arsenic, copper, and nitrates frequently present problems for private wells, data for
these compounds came from public water supply systems. In addition, extrapolation from
data presented in one state (New Hampshire) to the entire region presents a great uncertainty.
It should also be noted that lead, radon, and contaminants released from leaking
underground storage tanks are explicitly excluded from the analysis for this problem area in
order to conform to the definition for this specific problem area. In Region I, leaking
underground storage tanks have been identified as one of the major contributors to private
well contamination. Lead found in drinking water, primarily as a result of corrosion from
delivery piping systems, also causes concern.
Radionuclides other than radium are not analyzed. Uraninum data are lacking for this
analysis. The analysis of radionuclide data in drinking water, excluding radon and uranium, is
only a very small portion of the radionuclide problem.
47
-------
13. RCRA Waste Sites
Problem Area Definition
This category includes the risks posed by hazardous waste sites regulated under the
Resource Conservation and Recovery Act (RCRA). More specifically, it includes operating
and inactive RCRA landfills and surface impoundments, hazardous waste storage tanks,
hazardous waste burned in boilers and furnaces, hazardous waste incinerators, and associated
solid waste management units. Seepage and routine releases from these sources contaminate
soil, surface water, and ground water, and pollute the air. There is potential double-counting
with risks estimated in Hazardous/Toxic Air Pollutants (problem area #3), and Discharges to
Surface Waters (problem areas #7, #8, #9).
Summary/Abstract
The RCRA problem area poses relatively low individual cancer risk (1E-6) and represents
approximately 1 to 10 cancer cases per year. The cancer range derives from a rough estimate
of the possible impact of solid waste management units (SWMUs). The percentage of the
problem covered and the uncertainty are high. The uncertainty is associated with the use of
modeling results to estimate ambient concentrations. Non-cancer risks were estimated to be
low since no releases from sites were estimated to result in doses greater than the RfD.
Toxicity Assessment
Based on estimates of typical constituents of the hazardous wastes generated in Region I,
more than 100 compounds could be expected to be found at RCRA sites across six states. We
evaluated cancer and non-cancer endpoints for 18 of these compounds, including arsenic,
benzene, cadmium, carbon tetrachloride, chloroform, chromium, dichloromethane, mercury,
PCBs, toluene, vinyl chloride, and xylene. The endpoints of concern for these compounds
vary from cancer to a variety of systemic effects, including reproductive effects and
neuro-behavioral effects.
48
-------
Exposure Assessment
There are more than 5,200 hazardous waste generators in Region I and more than
228 treatment, storage, and disposal facilities (TSDs). Most of the TSDs in Region I are
storage facilities or treatment and storage facilities, with only one operating land disposal
facility. Most of the waste generated in the region is shipped to other parts of the United
States or to Canada for final disposal. Very little waste is imported into Region I for
handling. As a result, risks to the population in New England are primarily from tank and
drum storage, treatment, and waste previously disposed of in landfills.
Risk was assessed using the Regional Hazardous Waste Planning Model previously
developed by Temple, Barker & Sloane, Inc. (TBS) for EPA. The model was designed to
assess the relative costs and risks associated with various hazardous waste management
strategies (e.g., increased use of incineration). In estimating risk, the model uses
generator-specific data on quantities of waste handled, management practices, location,
average release algorithms, and data on exposure parameters. The waste data come from
validated biennial reports filed with each state by the generators. The waste constituent data
and release algorithms were developed by the Office of Solid Waste. The exposure
parameters (hydrogeological setting, temperatures, population locations, drinking-water
source type) were developed by TBS for each geographical area in the region where waste
handling takes place. These characterizations are rough in some cases, with each location
being classified into one of several possible characterizations for each exposure parameter.
For example, there are only three possible stream types to describe any given geographical
area.
Release and exposure pathways were evaluated for each waste-handling method. The
evaluations considered multiple exposure points, air concentrations at four separate distances
(from 1 to 19 kilometers) from the source, ground-water concentrations estimated at 60 meters
from the source, and surface-water concentrations estimated at 10,000 meters from the source
(surface waters are assumed to be treated before ingestion). The total population at risk varies
for each exposure pathway. In all cases, however, it is more than one million people
(category 4 in our scoring system).
Risk Characterization
Average individual cancer risk is on the order of 1E-6 across the entire region. Total
population cancer risk is low, approximately one case per year. The majority of this risk is
attributed to hazardous waste incineration. If the blending and burning of used oil in boilers
and furnaces had been evaluated for this problem area, this risk level would have been higher.
The Hazardous/Toxic Air Pollutants problem area evaluated the burning of used oil. No
releases from sites were estimated to result in doses greater than the RfD.
The consideration of SWMUs could raise this risk estimate. Based on an examination of
RCRA Facility Assessments for nine facilities in New England, more than 200 SWMUs or an
average of more than 20 SWMUs per site were found. These SWMUs were generally of two
types: tanks and drums or unlined landfills and lagoons. Further investigation found that the
materials in these SWMUs were generally consistent with current activities and wastes at the
site.
49
-------
Assuming that the waste types and annual volumes are roughly equivalent to the present
activities at the sites, we developed a very rough estimate of the impact of the SWMUs. The
drums and tanks currently regulated pose very minimal risk as estimated by our model; we
estimate that the SWMU drums and tanks probably pose a higher but still minimal risk. The
SWMUs that are unlined landfills and lagoons, however, may pose risks one to two orders of
magnitude higher than the risks posed by currently regulated RCRA landfills. This higher
risk is due primarily to the large number of landfill SWMUs. These SWMUs could therefore
add another 1 to 10 cancer cases per year to the risk estimated for this problem area.
The Public Health Risk Work Group estimated the percentage of the problem covered as
high (all RCRA sites, many pollutants, total waste volume, rough SWMU estimate) and the
degree of uncertainty as high. The high uncertainty derives from the use of modeled
concentrations rather than monitoring data.
50
-------
14. Superfund Waste Sites
Problem Area Definition
This problem area includes waste disposal sites that are regulated by Superfund.
Generally they are inactive and abandoned. It includes sites on the National Priorities List
(NPL), those deleted from the NPL, those that are candidates for the NPL, and any additional
sites that states may be addressing. These sites are typically characterized by the presence of
hazardous waste and the absence of any current manufacturing, treatment, storage or disposal
activity. These sites may contaminate ground and surface water, sediments, and soil, and may
pollute the air.
Summary/Abstract
Superfund sites present an increased individual cancer risk in the 1E-5 range and a
population risk of less than one case of cancer per year in Region I. Estimated doses are
rarely expected to exceed the RfD, and when they do they are generally on the order of
5 to 10 times the RfD. This assessment is based on a detailed analysis of the Endangerment
Assessments (EAs) prepared for 13 Superfund sites. The results from the 13 sites were
extrapolated to cover the approximately 1,570 sites in Region I that fit this problem category.
Both the portion of the problem covered and the uncertainty (due to the extrapolation of
results) are considered high.
Toxicity Assessment
The chemicals of concern vary from site to site depending on the particular activities or
wastes disposed at the site and the environmental and exposure factors at the site.
Approximately 25 Superfund risk assessments were reviewed, but only 13 were considered
comparable and therefore evaluated in depth. Documents were not considered comparable
primarily because mprimqm rather than average concentrations had been used in the EA
analysis. Of the 13 NPL sites examined in detail, 36 different chemicals were considered in
the Endangerment Assessments. The most frequently considered chemicals were vinyl
chloride, tetrachloroethylene, toluene, methylene chloride, benzene, PCBs, trichloroethylene,
and dichloroethylene. The driving pollutants in the health risk analysis were PCBs,
tetrachloroethylene, benzene, vinyl chloride, chloroform, bis-2-ethylhexyl phthalate,
benzo-a-pyrene, trichloroethylene, and arsenic. These pollutants occurred in concentrations
and exposure pathways that dominated other pollutant/concentration/ exposure combinations
at the 13 sites examined in detail. The Cancer Potency Factors (CPFs) for these pollutants
range from O.OS2 (mg/kg-day)"1 for tetrachloroethylene to 4.34 (mg/kg-day)"1 for PCBs. The
only pollutants for which an RfD was exceeded were dichloroethylene and
tetrachloroethylene. The health effects associated with these RfDs are hepatic lesions for
dichloroethylene and liver and kidney changes for tetrachloroethylene.
51
-------
Exposure Assessment
There are currently 59 sites on the NPL list in Region I (as of early June 1988) and a total
of approximately 1,700 sites that are either on the NPL list, are candidates for the NPL list, or
are being addressed by the six New England states. Approximately 130 sites require no
further action, thus leaving approximately 1,570 sites (59 of which are current NPL sites) that
fit this category. The sites represent a variety of activities (manufacturing, municipal waste
disposal, hazardous waste storage/recycling, etc.) and are located throughout the region in
both rural and urban areas.
For this analysis, we examined recently completed EAs for 13 Region I NPL sites. The
results from these sites were used to characterize the risks at all NPL sites and ultimately at all
1,570 sites in this problem area.
The risks estimated represent the risk posed to an "average exposed individual";
reasonable exposure pathways and mean concentrations, rather than maximum concentrations
reported in monitoring data, were used. We used the information in the EAs (e.g., doses,
concentrations) to estimate risks associated with reasonable exposure pathways, i.e., pathways
that could be expected to result in human exposure now or in the future and that had not
already been controlled. As a result, worst-case scenarios, such as future construction of a
residential well in the middle of a disposal site, were not included, but offsite exposures and
risks from trespassing were.
It is important to note that these pathways and risks are not necessarily those selected in
the EAs as representative of "current" or "future" risks, because the EAs often evaluate the
risk from developing the site for residential use. Also, we selected the highest risk pathway of
all the "reasonable" pathways evaluated at a site to represent the risk. That is, we did not add
risks from more than one pathway to arrive at an aggregated individual or population risk.
The exposure pathways examined at each site varied. Seven sites evaluated ground-water
ingestion as the pathway of concern; two examined dermal exposures to trespassing children
or teens; three examined inhalation exposures; and one examined the risks from fish ingestion.
The populations at risk at each site were estimated from information in the EAs. The
population is that associated with the particular pathway selected to represent the site. It was
often very roughly estimated. The exposed populations at the 13 sites evaluated varied from
25 for ground-water ingestion at one site to 1,500 for ground-water ingestion at another site.
-52
-------
Risk Characterization
In order to determine if our analysis of the 13 sites could be considered in any way to be
representative of the other 46 NPL sites, we examined two pieces of information. First we
computed the average Hazard Ranking System (HRS) score for all 59 NPL sites (45) and
compared it to the average HRS score for our sample (40). Although the HRS ranking is a
rough ranking, we felt we could conclude that our sample did not appear to include sites that
were significantly more or less risky than the total universe of NPL sites in Region I. We also
determined that the site type profile of our sample was not significantly different from the
59 NPL sites as a whole (see Table 14-1). As a result, we felt confident that our results would
not be radically different from results obtained for the population as a whole.
Table 14-1
Distribution of Supertund Site Typo
Sites Sample Region I NPL
Private Landfills
Municipal Landfills
Manufacturing Operations
Hazardous Waste
Storage/Recycling
Other
Total
5
3
2
2
1
13
15
8
12
11
13
59
After much discussion in the work group, it was decided that the other non-NPL sites
could be regarded as being of similar risk. There was no reason to assume that those sites
were any more or less risky then the present NPL sites. As a result, it was felt that the sites
examined could be used to represent all the sites categorized in this problem area.
The increased individual cancer risk across the 13 sites evaluated ranged from 1.3E-6 to
1.3E-2. The median value for the sites was approximately 1.3E-5. RfDs were exceeded at
two of the 13 sites with the Dose/RfD ratio varying between two and nine for
tetrachloroethylene and dichloroethylene.
Population risk was estimated by using the median number of people exposed at the
13 sites examined (250 people) and multiplying that figure by the number of sites (1,570) to
arrive at the population exposed (approximately 400,000). This number multiplied by the
median individual risk (1E-5) resulted in an estimated cancer risk for the problem area of four
cancer cases over a 70-year period, or less than one case per year.
53
-------
Using the non-cancer scoring system discussed elsewhere, the following scores were
assigned to this problem area: 1 for individual exposure ratio, 1 for population exposed to
non-carcinogens, and 1 for severity score.
The work group rated the uncertainty associated with these estimates as high, based
largely on the extrapolation from the small sample size to the large population. The portion of
the problem covered was rated as high by the group, based on the fact that numerous
chemicals were considered in the evaluation.
54
-------
15. Municipal Waste Sites
Problem Area Definition
This category consists of open and closed municipal landfills, municipal sludge and waste
incinerators, municipal surface impoundments, land application units, and land treatment
units. These sources can contaminate ground and surface water, the soil, and air. There is
potential double-counting of the risks from this category with those from Hazardous/Toxic
Air Pollutants (problem area #3), Discharges to Surface Waters (problem areas # 7, # 8, # 9),
and Other Ground-Water Contamination (problem area #19).
Summary/Abstract
The average individual risk associated with exposure to drinking water 600 meters from a
municipal landfill is estimated to be on the order of 1E-S. Population risk was not estimated.
The percentage of the problem covered is medium because approximately half of the
municipal facilities were evaluated. The air exposure risks were not estimated. The
uncertainty is high due to the use of modeled concentrations and few data points.
Toxicity Assessment
The contaminants from municipal landfills include a set of approximately 200. The group
evaluated for the risk evaluation includes vinyl chloride, arsenic, tetrachloroethane,
dichlormethane, carbon tetrachloride, antimony, and phenol These contaminants were
selected due to their concentrations in municipal landfill leachate, toxicity to humans,
regulatory limits under SDWA, and mobility and persistence in the subsurface. Five of the
contaminants are carcinogens. The non-cancer effects are neurotoxicity, cardiovascular
changes, and kidney and liver effects.
55
-------
Exposure Assessment
The problem definition in Region I includes:
operating landfills 739
closed landfills 1,443
sludge and waste incinerators 20
municipal surface impoundments 2,000
This analysis covers only the operating landfills. We have information on 58 of the 739
operating landfills. Of the landfills in Region 1,75 percent are categorized as small, receiving
less than 30 tons of waste per day; 23 percent are medium, with 30 to 50 tons of waste per
day; and 2 percent receive more than 500 tons per day. The tons of waste received per day,
along with the mean net precipitation, enabled us to calculate the leachate volumes for each
category of landfill. The leachate volume was then multiplied by the median leachate
concentration of each of the stressors (derived from the Subtitle D Regulatory Impact
Analysis).
To obtain the concentration of the contaminants in the ground water, EPA's Liner
Location Model was used. The model has several ground-water scenarios; flow field B, with
a ground-water velocity of 10 meters per year in the horizontal and 1 meter per year in the
vertical, was used most often. The flow field, mobility of the contaminants, and the
concentration of the leachate were combined to give the daily emissions of the contaminants
at a variety of distances. The 600-meter exposure distance was chosen because our database
showed the average distance to a private well to be 409 meters and the average distance to a
public well to be 680 meters.
The database of 58 active municipal landfills showed that 43 percent of the landfills had
wells within 1 mile. There were 246 wells in total and, assuming that three people drink
water from each well, 738 people were potentially exposed. Five percent of the landfills had
public wells within one mile, with five wells total We did not attempt to estimate the total
number of people potentially exposed from this water source because we could not determine
an average number of people per water supply.
Extrapolating from the 58 landfills to the 2,182 active and inactive landfills, we find that
21,471 wells may be within 1 mile of the municipal landfills. Therefore, potentially 64,413
people may be exposed.
56
-------
Risk Characterization
The health risk was characterized by converting the drinking water concentration
600 meters from each landfill into a dose. The dose was then multiplied by a CAG score for
the carcinogens or compared to the RfD for non-carcinogens. The individual risk associated
with each of the 58 landfills was computed by adding the risk associated with the five
carcinogens and by examining the highest resulting non-carcinogen dose to RfD ratio. No
landfills had dose-to-RfD ratios greater than one. The cancer risk associated with 60 percent
of the landfills was IE-10,20 percent had risk in the 1E-4 to 1E-5 range, and 20 percent had
risk in the 1E-3 to 1E-4 range. The average cancer risk is 1E-S.
The percentage of the problem covered is medium because approximately half of the
municipal facilities were evaluated. The air exposure risks from municipal incinerators and
gaseous emissions from landfills (H^S, methane, and vinyl chloride) were not estimated. The
uncertainty is high due to the use of modeled concentrations and few data points.
57
-------
16. Industrial Waste Sites
Problem Area Definition
This category covers industrial waste, including sludges handled in nonhazardous
industrial landfills, industrial surface impoundments, land application units, and land
treatment units subject to Subtitle D, along with numerous incinerators. Routine and
nonroutine releases, solid migration, and runoff may contribute particulates, toxics, biological
oxygen demand (BOD), PCBs, and nutrients to air, surface water, ground water, and soil.
Risks from the category could be double-counted with Other Ground-Water Contamination
(problem area #19), Discharges to Surface Waters (problem areas #7, #8, and #9), and
Hazardous/Toxic Air Pollutants (problem area #3).
Summary/Abstract
The Public Health Risk Work Group analyzed this problem area in relation to the
municipal waste sites due to lack of data on industrial waste sites. The work group reasoned
that the releases emanating from industrial waste sites may contain .similar contaminants but at
higher concentrations than releases from municipal sites. Therefore, we agreed that the risks
associated with industrial sites are also in the range of approximately 1E-S. Industrial sites
were ranked in the same group as RCRA, Superfund, and Municipal Waste Sites for cancer
and non-cancer risks. Refer to problem area #15, Municipal Waste Sites, for additional
information.
58
-------
17. Accidental Releases
Problem Area Definition
The uncontrolled accidental release of hazardous substances with a potential for individual
or group overexposure by any route with potential serious adverse effects is identified as the
problem area. Such releases-may be from stationary sources or mobile sources.
Summary/Abstract
Acute, serious, localized, non-cancer effects are considered to be of primary importance in
the defined problem area. The frequency of such effects and the numbers of individuals likely
to be affected in such events are both believed to be relatively small. In 1985,
approximately 800 reported incidents occurred throughout the six states in Region I.
The hazards of the major chemicals of concern are high, especially when they are in the
gaseous or vapor phases. Chlorine (gas) features in many reported releases from stationary
sources such as paper-making facilities; ammonia gas and hydrogen sulphide gas are other
common chemicals of concern. All three substances are extremely toxic in any form.
The health effects associated mostly with acute releases of both inorganic and organic
substances are both systemic and local. With few exceptions (notably, asbestos and
chlorinated organic compounds), cancer is not associated with an acute release.
The most common pathway of exposure involving acute hazardous releases of toxic gases
or vapors is inhalation; here, dermal and ingestion routes are generally quite secondary.
However, exceptions exist (notably, high concentrations of acetonitrile, acrylamide, phenolic
compounds, aniline and some other primary amines with significant vapor pressures at
ambient temperatures).
Exposures with the potential for acute adverse health effects, in the case of accidental
releases, are believed to be generally restricted. Individuals in the immediate vicinity of the
events or who are present within relatively small areas downwind or to the sides of die release
are at the highest risk; Le., the populations at greatest immediate risk, even in high-volume
bursts, are best characterized as "immediate area/near neighbor populations."
The risks to those who are acutely exposed to nearby releases may be best characterized
qualitatively; no satisfactory statistical basis is evident for quantitatively characterizing such
non-cancer risks.
59
-------
Risks to populations at large, with respect to non-cancer effects, are believed to be
generally small. Attenuation of concentrations of toxic releases is oftentimes claimed to be
rapid, in terms of both time and distance, by officials reporting on affected sources. This
claim, however, is by no means axiomatic. Topography and meteorological conditions can
profoundly affect the extent of potential serious effects associated with some releases of
highly toxic substances. (Bhopal is a case in point In this disaster, meteorological conditions
coupled with high population densities in areas surrounding the chemical plant played a large
pan in the catastrophe.) In such situations, however, it can be very difficult to quantify risk
with any generally acceptable level of certainty.
Toxicity Assessment
The common major chemicals of concern in this problem area are basic commercial
chemicals: strong mineral acids (sulphuric, hydrochloric, and nitric acids), anhydrous
ammonia, chlorine, strong mineral bases (sodium hydroxide and aqueous ammonia), and
commercial solvents with relatively low boiling points and high flammability (methyl alcohol,
toluene). Most of these common chemicals have appreciable toxicity and/or are potentially
explosive. In terms of personal risk, inhalational hazards appear to outweigh individual risk
of fatality by fire, based on the generally limited database available for consideration here.
Exposure Assessment
Exposures to excessive levels of airborne concentration of major chemicals of concern is
believed to be the more likely cause of acute effect(s) in the event of an accidental release.
The major chemicals of concern have assigned ceiling limits but do not have "skin" notations
in the "Threshold Limit Values" established by the American Conference of Governmental
Industrial Scientists.
Risk Characterization
Risks associated with accidental releases are believed to be essentially non-cancer type
end-points. Cancer risk with most of the substances that might be released (even if the
substance has a high cancer potency score) is believed to be very low with a single episode.
In this case, there appears to be no reliable method to quantify cancer risk.
The uncertainties in the characterizations and qualitative assessments provided here and
summarily reported on are believed to be small.
Acute episodes of uncontrolled releases of hazardous chemicals are extensively reported
upon and receive media attention, even when no personal injuries are associated with the
release. Frequency data (e.g., the number of chlorine releases per year) are known with
reasonable certainty. However, the numbers of the individuals within known distances of the
release sources and who are actually adversely affected are not known at mis time with any
acceptable degree of certainty. For this reason, risks associated with acute releases are
probably best described qualitatively.
60
-------
18. Releases from Storage Tanks
Problem Area Definition
This category includes releases of petroleum products or other chemicals from tanks that
are above, on, or under the ground, tanks owned by fanners, and the fuel-oil tanks of
homeowners. Stored products include motor fuels, heating oils, solvents, pesticides, and
lubricants that can contaminate ground water with such toxics as benzene, toluene, and
xylene. The primary environmental hazard is contamination of ground water and soil.
Summary/Abstract
The releases from storage tanks are likely to pose low individual cancer risk and possible
narcotic effects due to inhalation of gasoline and fuel oil through seepage into homes and
during showering. The percentage of the problem covered is low due to the evaluation of
motor fuel tanks only. The uncertainty is high because of lack of information regarding
concentrations, toxicity, numbers of leaking tanks, and people exposed.
Toxicity Assessment
The risk evaluation focused on petroleum releases. There are literally hundreds of
constituents in fuel oil and gasoline. The analysis focused on hydrocarbons, benzene, toluene,
xylene, and ethylbenzene. We did not directly evaluate cancer risk or non-cancer endpoints
from these compounds due to the lack of data. The major health effects follow inhalation.
The effects include narcotic effects, polyneuropathy, general anesthetic effects, and depressant
effects on the central nervous system. Some of the organic compounds, such as benzene, are
carcinogens.
Exposure Assessment
The number of tanks in Region I is estimated at 758,500, with 143,500 being regulated
motor fuel and chemical tanks. The remainder of the tanks are residential and non-residential
heating oil tanks and farm heating/motor fuel tanks. The amount of motor fuel and heating oil
stored in tanks in Region I is about equal, although the number of heating oil tanks is much
greater.
61
-------
The Underground Storage Tank (UST) program estimates that 10 percent of all tanks at
gasoline stations are leaking. Applying that estimate to the total number of tanks in New
England, there would be 75,850 leaking tanks. The number of leaking motor fuel tanks
would be 14,350. However, we used the estimate of 10,202 because this figure is derived
from the tanks reported in the state UST programs. Case studies from Maine showed that
hydrocarbon contamination in drinking-water wells ranged from approximately 10 ppb to
40,000 ppb. Benzene levels were 0 to 1,400 ppb, toluene 0 to 7,500 ppb, ethylbenzene
0 to 1,800 ppb, and xylene 0 to 8,800 ppb.
Other potential pathways include inhalation through seepage into homes and showering.
Using the exposure levels from the Maine case studies may overestimate the presence of
contamination and concentration levels in the drinking water. In Maine, the gasoline may
travel farther and faster due to bedrock fissures acting as conduits. Other New England states
have less bedrock geology.
Again using data from Maine, we learned that there are 155 leaking tanks resulting in:
Private wells - polluted 219
Private wells - threatened 174
Public wells - polluted 2
Public wells -threatened 5
Extrapolating from the Maine data, we found that there would be 14,414 total
contaminated wells in New England, based on the estimated 10,202 leaking tanks. There
would be 25,867 threatened and contaminated private wells. Multiplying the number of wells
by three people per well, we find 43,242 people potentially exposed to contaminated well
water and 77,601 people potentially using water from threatened and contaminated wells. We
did not calculate the number of people potentially exposed from public wells because we
assumed that the population estimate would still fall into the 3 category in our scoring system.
Our analysis did not focus on home heating oil tanks or farm tanks. Had we analyzed the
health risks from these tanks, the number of potentially exposed people would likely have at
least doubled.
Risk Characterization
Due to the lack of contaminant-specific data, we did not estimate risk for individuals and
the population. The work group estimated that the cancer risk would fall into category 2,
given the presence of known carcinogens but at unknown levels. Additionally, we are not
aware of any known concentration or range of concentrations where an "average" individual
62
-------
would cease drinking water contaminated with gasoline due to taste and odor considerations.
For non-cancer rankings, the group estimated that the severity ranking would be 3 due to the
endpoints mentioned above. The population score was 3 and the potency score was not
estimated. Overall, the group felt that the non-cancer risks were in Category 3, and are
therefore greater than the risk associated with the various waste programs.
The percentage of the problem covered is low due to the evaluation of motor fuel tanks
only. The uncertainty is high because of lack of information regarding concentrations,
toxicity, numbers of leaking tanks, and population exposed.
63
-------
19. Other Ground-Water Contamination
Problem Area Definition
An extensive variety of point and nonpoint sources of pollution not addressed in other
problem areas also contaminate ground water. For the purposes of this project, only septic
systems, road de-icing salts, Class V underground injection wells, and the leaching of
agricultural pesticides and fertilizers were assessed in this problem area due to limited
databases, resources, and time. Impacts from underground storage tanks, hazardous waste
sites, and landfills are covered in other problem areas for this project
The list of possible contaminants is even more extensive than that of possible sources.
Possible contaminants include microbes, nitrates, sodium, chloride, pesticides, and, due to the
inclusion of class V underground injection wells, potentially any waste fluid produced by
various industries, utilities, and commercial ventures, including toxic organic and inorganic
chemicals, heavy metals, and oil and petroleum products.
Summary/Abstract
The greatest potential risk from this problem area is non-carcinogenic, that assessment
being based primarily on bacterial and viral contamination from septic tanks and cesspools.
The smaller, carcinogenic risk of this problem area is also thought to be driven by septic tanks
and cesspools. This was the only source in this problem area for which a quantitative
assessment was attempted.
Septic tanks drive the risk in this area primarily because of their widespread, densely
located occurrence in much of the region and a high estimated population at risk (60 percent
of the 2 million private well users in New England). The other sources considered also pose
some combination of carcinogenic and non-carcinogenic risks to a smaller portion of the
2 million private well users. The assessment is highly qualitative because of a lack of data.
The data gap is primarily a function of the population assessed, because we considered only
private wells and the water-quality data for these wells are lacking in both space and time.
Often, data are reported as well-contamination incidents rather than as actual concentration
data. Most private wells in the region are probably at risk from at least one of the sources
covered here. The uncertainty in this assessment is high, and the extent of the problem
covered is low.
64
-------
Toxicity Assessment
Major ground-water pollutants from septic tasks and cesspools include microbes
(e.g., pathogenic bacteria and viruses), nitrates, chloride, methylene chloride (a.k.a.
dichloromethane) and 1,1,1-trichloroethane. Major pollutants from road de-icing salts include
sodium and chloride. No "priority" pollutants were assigned to the Class V underground
injection wells source because of the variety of well types and lack of data. The pollutants
from the wells include, potentially, any waste fluid produced by various industries, utilities,
and commercial ventures, as well as septic system pollutants and agricultural chemicals.
Major pollutants found in ground water from agricultural pesticides and fertilizers include
nitrates, aldicarb, EDB, alachlor, atrazine, oxyamyl, dinoseb, carbofuran, 1,2- and
1,3-dichloropropane.
Table 19-1 identifies the kinds of microorganisms found in domestic wastewater and the
associated disease endpoints that drive the non-carcinogenic risk from contamination by septic
tanks and cesspools.
Coliform bacteria are used as indicators of contamination by human and/or animal waste
because they are routinely shed in the feces. The primary drinking-water standard for
coliform bacteria is less than one per 100 ml of water, the absolute level allowable varying
with sampling frequency. The assumption is that if coliform are present, then any other
microorganism could be, too, and it may take only one microorganism to produce the
resultant disease in a human.
Infants are primarily the population at risk from high nitrate levels in drinking water
because they do not have the acids in their digestive system to prevent the growth of bacteria
that convert nitrates to nitrites. Nitrites reduce the oxygen-carrying capacity of the blood and
can cause death by methemoglobinemia, a type of "suffocation."
The carcinogenic potential of contamination by septic tanks and cesspools is due to toxic
chemicals that can be released into such systems by household products and septic tank
cleaners. Table 19-2 lists the predicted priority pollutants that can occur in household
wastewater.
The primary pollutant chosen to estimate some of the carcinogenic potential of ground
water contaminated by septage was methylene chloride (a.k.a. dichloromethane). This was
selected because it was the only carcinogenic organic chemical noted in the literature that was
cited as being found frequently. It was noted as a septic leachate pollutant by Rhode Island in
its 1988 Water Quality Assessment report and was the same chemical evaluated for this
source of ground-water contamination in the national comparative risk assessment project
Methylene chloride has a carcinogenic potency factor of .0075 (mg/kg/day)'1 for oral
ingestion and is currently classified as a probable human carcinogen.
65
-------
Table 19-1
Pathogenic Microorganisms In
Domestic Wastewater
(USEPA, June 1987)
Microorganism
Dlsease(s) Caused
Bacteria
Salmonella species
Shlgella
Yersinia
Mycobacterium
Leptosplra
Campylobacterjejuni
Pathogenic coliforms
Yersinia enterocotttica
Pseudomonas
Klebsiella
Serrate
Viruses
polioviruses
hepatitis A
echoviruses
coxsackieviruses
Norwalk and Norwalk-like viruses
rotaviruses
adenoviruses
Parasites
Entamoeba hlstolytlca
Glardla lambia
Balantidium col!
Ascarisova
Trichuris
Enterobius vermicularis
Cestode ova
Coccidia
typhoid, paratyphoid,
gastroenteritis
bacillary dysentery
gastroenteritis
tuberculosis
leptospirosis
gastroenteritis
gastroenteritis, urinary
tract infections
gastroenteritis
respiratory and bum
infections, diarrhea
pneumonia, bronchitis
respiratory and urinary
tract infections,
summer diarrhea
poliomyelitis
infectious hepatitis
respiratory disease,
aseptic meningitis,
diarrhea, fever
respiratory disease,
fever, aseptic meningitis,
myocarditis
gastroenteritis
respiratory disease,
eye infections
amoebic dysentery
giardiasis ("backpacker's
diarrhea")
dysentery, gastroenteritis
pneumonitis, intestinal
and nervous system
disorders
chronic gastroenteritis
enterobiasis
chronic gastroenteritis
diarrhea, toxoplasmosis
66
-------
Table 19-2
Predicted Priority Pollutants
In Household Wastewater
(ISDS Task Force, 1986)
Organlcs Inorganics
benzene arsenic
phenol cadmium
2,4,6-trichlorophenol chromium
2-chlorophenol copper
1,2-dichlorobenzene lead
1,4-dichlorobenzene mercury
1,1,1-trichloroethane zinc
naphthalene antimony
toluene silver
diethylphthalate
dimethylphthalate
trtehloroethylene
aldrin
dieldrin
With road de-icing salts, primary concern is for sodium as a pollutant Excessive sodium
is unhealthy in any diet, but some segments of the population require low sodium diets for'
health reasons. High sodium levels stress cardiovascular, liver, and kidney conditions and can
affect blood pressure. High sodium levels also may be related to hypertension.
The agricultural pesticides listed as primary pollutants for this evaluation are the ones
most commonly found in ground water. They have varying toxicity and carcinogenic
potentials. Although many have been well studied regarding health effects because of
concerns about food residues, their health effects due to ground-water ingestion are less
understood owing to long-term, low-level exposure scenarios and a complex array of physical
and chemical factors determining their fate and transport to and in ground water. The only
concentration data obtained for this assessment were for aldicarb contamination. Aldicarb
(trade name TEMIK) can be adsorbed through all routes of exposure, with ingestion being the
most significant It is acutely toxic, interfering with the transmission of nerve impulses. It
has not been shown to be carcinogenic, mutagenic, or to cause birth defects. A study in
Wisconsin referenced by the Commonwealth of Massachusetts (1986) showed that aldicarb
had an effect on the immune system at concentrations less than 10 ppb, which is the
Massachusetts interim guideline standard.
67
-------
Fertilizers are a source of nitrates, the health effects of which are discussed on page 65.
Because of the wide variety of pollutants possible from Class V underground injection
wells, no specific health concerns were noted for pollution from this source. For the septic
and agricultural wells, other discussions in this section can be referenced.
Exposure Assessment
Sources
Obtaining data on the types, numbers, sizes, and locations of the major sources in this
problem area requires documented inventory by the states and their local communities.
Whereas a town might have one landfill, it could have hundreds of septic systems, many miles
of roads that are salted in the winter, numerous unrecognized Class V underground injection
wells discharging hazardous and nonhazardous wastes, and some areal extent of agricultural
fields where pesticides and fertilizers are applied. The most readily available statistics about
the separate sources appear below.
Septic tanks and cesspools include individual, on-site subsurface disposal systems serving
fewer than 20 people. Of all ground-water pollution sources (i.e., not just those in this
problem area), septic tanks and cesspools rank the highest in volume discharged directly to
soils and are the most frequently reported sources of ground-water contamination (Canter and
Knox, 1985). The estimated number of septic systems and a calculation of statewide density
of systems per square mile, from 1980 Census of Housing data, appear in Table 19-3.
The total number of septic tanks and cesspools for the region in 1980 was about
1,480,500. Recent estimates by some of the New England states show an increase since 1980:
26 percent for Maine (now about 250,000 systems), more than 9 percent for Massachusetts
(now more than 600,000 systems) and 28 percent for Rhode Island (now about 143,900
systems). It is probably reasonable to assume that there are about 1.75 million septic tanks and
cesspools in New England. Table 19-4 indicates some county-level statistics, also based on
1980 Census data, which were calculated assuming a uniform distribution of systems within
the county.
Septic tanks and cesspools are located throughout the region in unsewered areas. An
assumption that they occur only in very rural settings would be erroneous, as can be seen by
examination of the counties listed in Table 19-4, which are among the most populous in the
region. If the design life of these systems is 10 to 15 years, we can expect that we are seeing
the beginning signs of a problem that could only get worse as development in New England
increases, because many of the septic tanks and cesspools in New England were installed in
the 1960s. It is estimated that only 40 percent of all septic systems function properly, and
areas with more than 40 systems per square mile can be considered to have potential
contamination problems (Canter and Knox, 1985). Reference to Tables 19-3 and 19-4 shows
that the more populous southern New England area have double and even triple this density in
some counties.
68
-------
Table 19-3
1980 Septic Tanks and Cesspools
State Total Number of Septic Systems (1980) Systems Per Square Mile
Connecticut 357,446 73.4
Maine 198,629 6.4
Massachusetts 550,629 70.4
New Hampshire 161,386 17.9
Rhode Island 112,663 106.8
Vermont 99,752 10.8
Source: U.S. EPA, January, 1987.
Table 19-4
1980 Septic Tanks and Cesspools/Density
County, State Systems Per Square Mile
Fairfield, CT 79-158
Hartford, CT 68-135
New Haven, CT 82-164
Bristol, MA 90-180
Middlesex, MA 61-122
Norfolk, MA 127-254
Plymouth, MA 76-153
Worcester, MA 33- 66
Source: Canter and Knox, 1985.
69
-------
Massachusetts, with more than 600,000 systems and noted surface water problems,
appears to be at the greatest risk. On Cape Cod alone it has been estimated that approximately
63.8 million gallons per year (mgy) of septage is generated and only 31 percent treated
effectively, leaving about 44 mgy to discharge to a highly vulnerable environment (CCAMP,
1988). If an estimated 1.75 million systems discharge 45 gallons per day (gpd) per person,
and we assume three persons per system, then 86 billion gallons of septage per year is created
and potentially 51.6 billion gallons of partially treated sewage are discharging annually in
New England.
Examination of the 1988 Water Quality Reports for the New England states gives some
indication in a few states of the possible extent of septic problems. Failing septic systems are
noted as a major contributor to the nonattainment status of 27 of 32 drainage basins in
Massachusetts and as a minor contributor to two other basins. In New Hampshire it has been
noted that on-site wastewater systems had moderate to minor impacts on 92.2 miles of rivers,
and in Maine approximately 31.3 percent of ground-water nonattainment areas resulting from
nonpoint sources (or 91.3 square miles) are estimated to be due to septic systems.
The road de-icing salts category includes the application and storage of de-icing salts and
sand/salt mixtures. Table 19-5 displays various types of road salt statistics available from the
states and other sources. Using the tonnage per year figures for some of the states, an
estimation can be made that 750,000 tons of salt are applied annually to major New England
roads. In Maine it is estimated that a total of 17.9 percent of ground-water nonattainment
areas resulting from nonpoint sources are due to road salt Plumes from 700 uncovered
storage piles account for 3.8 percent (or 10.9 square miles) and road salt applications account
for 14.1 percent (or 41 square miles). A Rhode Island study (FHA, 1981) showed high
concentrations of sodium and chloride thousands of feet downgradient from salt storage piles.
For the Class V underground injection wells, a diverse array of source types and an
alarmingly large potential for unrecognized ground-water contamination are included in the
category. By definition, a Class V well injects nonhazardous fluids into or above
underground sources of drinking water. However, many of these wells, by nature of design
and function, directly discharge an array of toxic pollution sources above or into ground
water. If these wells are injecting hazardous substances they should be classified as Class IV
wells, which have been banned nationally, and thus would have to be shut down or have their
discharge modified and regulated by permit. Table 19-6 is the May 1987 Class V well
inventory for New England as reported to Congress. The scatter and various weights of
numbers in Table 19-6 suggest an incomplete inventory of Class V wells in New England and
illustrate the variation in targeting priorities among the states with respect to the inventory of
specific types of wells. More recent figures available from a EPA Federal Underground
Injection Control Reporting System (FURS) database report, dated February 5,1988, indicate
a total of 449 wells, and, according to the states' files on data not yet entered into FURS, there
are 729 Class V wells in New England. The discrepancies are partially due to a new system
of well codes for reporting inventories and are being resolved. Table 19-7 lists what are
classified as Class V wells under the new reporting system. Note that wells involving
disposal of septage are for systems serving 20 persons or more.
70
-------
State
Table 19-5
Annual Road De-Icing Salts
Tons TWO'Lane State Pounds per
Per Miles and/or Lane Mile Per
Year of Road Local Road Storm Event
Storage Covered
Capacity ?
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
mm
55,000
259,367
139,434
—
90,000
mm
22,000
—
8,571
—
3,000
mm
both
state
—
-
state
160
160
300
250
150-300
250
„
700 piles
—
123,305 tons
91 piles
••
.
no
—
yes
—
~~
Table 19-6
New England Class V Well Inventory
(Note: refer to Table 19-7 for well codes)
State Total 502 5D3 5D4 5A7 5W11 5W12 5A19 5W20 5X28 5R21 5X26
CT
ME
MA
NH
Rl
VT
Totals
84
15
132
38
80
15
364
3
—
19
—
_
—
~22
^m
—
—
3
—
—
~3
12
_
10
16 2
„
_ _
1e "24
62
—
27
—
8
—
~97
mm
~
72
—
_
—
~72
mm
—
3
3
8
—
—4
6
15
1
13
59
5
~99
1
—
~
—
3
10
-14
m m ^m
_
_ _
1
2
—
— ~2
Source: U.S. EPA, May 1987.
71
-------
Table 19-7
Types of Class V Wells
Well Code
Well Description
5W9
5W10
5W11
5W31
5W32
5W12
5A19
5W20
5X2B
5F1
5D2
5D3
5D4
5G30
5A5
5A6
5A7
SAB
5R21
5B22
5S23
5X13
5X14
5X15
5X16
5X17
5X18
5N24
5X25
5X26
5X29
5X27
Untreated sewage waste disposal wells
Cesspools
Septic systems (undifferentiated disposal method)
Septic systems (well disposal method)
Septic systems (drainfield disposal method)
Domestic WWTP effluent disposal wells
Cooling water return flow wells
Industrial process water and waste disposal wells
Automobile service station disposal wells
Agricultural drainage wells
Storm water drainage wells
Improved sinkholes
Industrial drainage wells
Special drainage wells
Electric power reinfection wells
Direct heat reinfection wells
Heat pump/air conditioning return flow wells
Ground-water aquaculture return flow wells
Aquifer recharge wells
Saline water intrusion barrier wells
Subsidence control wells
Mining, sands, or other backfill wells
Solution mining wells
In-situ fossil fuel recovery wells
Spent-brine return flow wells
Air scrubber waste disposal wells
Water softener regeneration brine disposal wells
Radioactive waste disposal wells
Experimental technology wells
Aquifer remediation wells
Abandoned drinking water wells
Other wells
72
-------
Potential sources of agricultural pesticide and fertilizer contamination include land
application, disposal, spills and leaks by manufacturers and formulators, dealers, and
industrial, agricultural, and domestic users. For this assessment, only regional agricultural
applications of pesticides were estimated.
A study extrapolating applications per year of 12 pesticides, based on 1979 and 1982
national data, indicates that 2.04 million pounds per year are applied in New England, or
about 1,020 tons per year. This estimate does not include homeowner use, usage in forests,
rights of way, golf courses, and other nonagricultural uses (RFF, 1985). The pesticides that
generally have a higher potential to leach into ground water are those with soil half-lives
greater than two to three weeks (USEPA, Feb. 1988). Probably less than 1 percent of the
estimated 1,020 tons per year of agriculturally applied pesticides leach into ground water.
Fertilizers are a source of nitrates and leach into ground water much more readily than
pesticides. In Rhode Island, a total of 22,849 tons of fertilizer were used last year, and only
9 percent of that was for agricultural use, the rest being for lawn and golf course applications.
The only exposure pathway considered for all sources was ingestion of contaminated
ground water via private drinking-water supplies. Contamination of public drinking-water
supplies was presumed to be incorporated in the Drinking Water problem area (#12).
Estimates of Population Exposures
Septic Tanks and Cesspools
Examination of Tables 19-1 and 19-2 indicates that a large range of health effects could
result from excessive concentrations of any one of, or a combination of, the contaminants in
these lists. In order to evaluate the risk of nitrate contamination, we would need concentration
data and an estimate of the infant population using private well water. We have chosen to
evaluate microorganism contamination for the non-carcinogenic risk and only one toxic
chemical for the carcinogenic risk of septic tanks and cesspools.
An EPA report on septic tanks (June 1987) presented data showing that nationally,
between 1971 and 1979,63 percent of cases of illness caused by untreated and contaminated
ground water were caused by the overflow of, or leachate from, septic tanks or cesspools. The
report further states that "septic tanks represent a significant threat not only to preserving the
potability of ground water, but to human health." The report cites several case studies for
which viruses and bacteria were found in ground water from the point of discharge to as far as
175 feet away in areas with shallow water tables. The report concludes that there is potential
for microorganisms to travel great distances and certainly to reach private wells within 100 to
200 feet from the point of discharge. Several other case studies of disease outbreaks
attributable to septic tanks note acute gastroenteritis and hepatitis A as frequently occurring
health endpoints. These diseases were also noted as the most common in Rhode Island's
73
-------
1988 Water Quality Assessment report. A survey of private well contamination in
Massachusetts (Commonwealth of MA, 1988) states that although "exact figures on sodium,
bacteria, and nitrate levels were not obtained as a result of this survey, it is likely that either
sodium, bacteria, or nitrate has affected a private well in every municipality in the
Commonwealth." The true extent of disease due to bacterial or viral contamination of ground
water is unknown for several reasons. Private wells are not regularly monitored; coliform
bacteria levels may not correspond directly to levels of other microorganisms, especially some
small viruses that can travel farther, and there is an underreporting of diseases or a lack of
awareness that contaminated ground water is at cause.
Several documents were referred to regarding the toxic chemical pollution potential of
septic tanks and cesspools. Briefly, if toxic chemicals are present in domestic wastewater they
generally will find their way into the ground water because septic systems are not designed to
remove these chemicals. Also, most septic tank cleaners are up to 100 percent solvents by
weight and tend to do more damage than good to the septic tank by killing the naturally
occurring microorganisms that help treat the septage. A Rhode Island Individual Sewage
Disposal Systems (ISDS) Task Force report (ISDS, 1986) references a case study where a
sampling of septic tank effluent revealed 40 to 50 compounds at >1 ppb, five of which were
identified as EPA priority pollutants. Another referenced study found 1 to 10 ppb
concentrations of organic chemicals in the ground water near septic systems. For this
evaluation we assumed an average exposure concentration for methylene chloride of 1 ppb.
However, it should be kept in mind that if a septic system is malfunctioning such that a well is
being contaminated, then it is most likely that a variety of organic and inorganic chemicals
will be present in varying concentrations.
The population at risk from ground water contaminated by septic tanks and cesspools is
estimated at 1.2 million people. Approximately 2 million people use private wells in the New
England region. If there are an estimated 1.75 million septic systems serving two to three
people on average, then septic systems serve 3.5 million to 5 million people, many more than
private wells do. We cannot assume that every household on a septic system is on a private
well, but we can assume that every household with a private well is on a septic system. This
leads to a predicament: how do we know if the estimated number of malfunctioning tanks
(60 percent of the total) serves households on private wells? There is no easy way to
determine this; thus we estimated that 60 percent of the households served by a private well
have a malfunctioning septic system, which would place 1.2 million people at risk of drinking
contaminated ground water.
Road De-Icing Salts
Private wells are being contaminated by salt storage piles and road salt applications,
although most states are now covering their storage piles. Limited research reveals
well-contamination and well-closing statistics that are not conducive to a quantitative risk
assessment; that is, average concentration data are needed, rather than numbers of drinking
water violations, to conduct a quantitative risk assessment. The following discussion of
sodium and chloride contamination of private wells qualitatively indicates the extent of the
problem. The statistics are based on a sodium standard of 20 mg/1 and a chloride standard of
250 mg/1. (Although chloride is not a health problem, it is used as a contamination indicator
by those states with naturally occurring high sodium levels.)
74
-------
Vermont estimates that seven wells close each year due to excessive sodium and chloride
levels. New Hampshire replaces an average of 36 wells per year due to excessive chloride
levels. Both states note that it is primarily shallow, dug, or driven wells that are affected. In
Maine, 66 private wells, mostly drilled wells, were contaminated by chloride recently (this is
probably not an annual number). Connecticut notes that 18 wells were'contaminated by
sodium between 1986-1987, bringing a recent total to 61 wells. Rhode Island had private
well contamination surrounding eight of their storage sites, with sodium levels as high as
300mg/l.
The population at risk from road salts would be that portion of the approximately
2 million private well users who are on a restricted salt diet and whose well has been
contaminated by sodium levels greater than 20 mg/1.
Class V Underground Injection Wells
We did not cite any data regarding the contamination of private wells in New England
specifically due to Class V wells. It is presumed that this source of contamination may affect
those wells contaminated by an "unknown source." Once the states begin to fully inventory
and locate Class V wells, especially with regard to their proximity to ground-water wells,
private well testing may better identify this source area's contamination potential.
The population at risk may be any portion of the approximately 2 million private
ground-water well users in Region I, although the actual total number depends on the location
of the wells relative to points of ground-water use.
Agricultural Pesticides and Fertilizers
Maine, Massachusetts, Rhode Island, and Connecticut have the highest potential of the
New England states for ground-water contamination by pesticides. It is thought that, in
general, much less than 1 percent of the agriculturally applied pesticides leach into ground
water, although this would probably vary considerably depending on soil and other
conditions. Since pesticides are very expensive to test for, citizens are reluctant to voluntarily
test their wells, and thus the extent of contamination is unknown. Private well testing by the
states, when funding is available, is usually targeted for high risk agricultural areas.
Rhode Island has texted extensively for aldicarb in the southern area of the state. Total
aldicarb was positive in 457 of 2,952 samples (15 percent) taken from 1,136 wells, with
145 samples (or 5 percent) above the health advisory of 42 ppb. Examination of preliminary
results (not yet published) shows a mean concentration of 6 ppb for all positive samples and a
mean concentration of 52 ppb for the 5 percent above the 42 ppb health advisory level.
Connecticut noted 134 wells as being contaminated by pesticides between 1986-1987,
bringing the total of well contamination incidents to 417. Massachusetts found 77 wells in
eight communities contaminated by chemical pesticides. A private well survey, testing wells
in 27 Massachusetts communities for eight pesticides, found 61 wells above state interim
75
-------
guidelines with a total of 155 wells showing some detectable concentration. Of the 61 wells
over the standard, 24 were for EDB, 33 for 1,2-D, 22 for aldicarb, 1 for carbofuran, 1 for
alachlor, and 2 for dinoseb. The three-year study was conducted from 1983 to 1986. A 1987
survey found six wells in one community over the standards: one for alachlor, four for
atrazine, and one for metolachlor.
The population at risk from pesticide-contaminated ground water would be that portion of
the estimated 2 million private ground-water well users living in the vicinity of agricultural
areas.
Risk Characterization
It was very difficult to determine whether a quantitative risk characterization of this
problem area was feasible. Initially, the work group felt that an improvement could be made
on the results of the national comparative risk project, especially where the only estimate of
carcinogenic potential for a list of sources even longer than that considered here was based on
one septic tank contaminant However, a lack of data left us in the same predicament. The
main risk characterization attempt was to improve the national project's qualitative discussion
of the potential sources covered and to determine the possible extent of the problems in New
England. Since the exposed population considered was some portion of the estimated
2 million people using domestic ground-water wells, and water quality data for these wells are
incomplete, a qualitative assessment was generally all the work group could do.
The only attempts to quantify some of the risk for this problem area concern septic tanks
for both the non-carcinogenic and the carcinogenic potential. We felt that the risk of disease
by bacterial and viral contamination facing an estimated 60 percent of the private well user
population would drive the non-carcinogenic risk. Other non-carcinogenic risk is posed by
the organic and inorganic chemicals commonly occurring in domestic wastewater, by sodium
from road de-icing salts, by the potential contaminants from Class V underground injection
wells, and by agricultural chemicals, although the exposed population for these sources is
probably much lower than that exposed to microorganism contamination.
The carcinogenic potential of this category was nearly impossible to estimate. An attempt
was made to quantify risks from methylene chloride, also known as dichloromethane, one of
many possible toxic organic chemicals commonly found in septic tank leachate. The
estimated individual risk posed by methylene chloride is 2E-7, a lifetime population cancer
risk of less than one cancer case for an estimated 1.2 million people and an annual population
cancer risk of much less than one cancer case (based on a 70-year lifespan). The estimate was
based on an average concentration of 1 ppb, which was the range noted in a few case studies,
and not on actual, regional concentration data. If more concentration data were available on
the occurrence in ground water of the organic chemicals from all of the sources being assessed
in this problem area, the cancer case number would most likely be higher.
The uncertainty in the ranking is high, based on the qualitative nature of the analysis
compared to other problem areas in the risk evaluation project. The percentage of the
problem area covered is low.
76
-------
Bibliography
Canter and Knox. Septic Tank System Effects on Ground Water Quality. Chesea, Michigan:
Lewis Publishers, Inc., 1985.
Commonwealth of Massachusetts. Contamination in Municipal Water Supplies. Boston,
Massachusetts: Special Legislative Commission on Water Supply, 1986.
•
Commonwealth of Massachusetts. Private Well Contamination in Massachusetts: Sources,
Responses and Needs. Boston, Massachusetts: Special Legislative Commission on Water
Supply, 1988.
Federal Highway Administration (FHA). A Study of the Effects of Salt Storage Practices on
Surface and Ground Water Quality in Rhode Island. NTIS report no.
FHWA-RI-RD-80-01,1981.
Individual Sewage Disposal Systems flSDS) Task Force. "Contribution of ISDS to
Ground-Water Contamination by Non-Conventional Pollutants." Draft Report, 1986.
Massachusetts Audubon Society. Road Salt and Groundwater Protection. Groundwater
Information Flyer #9, Community Groundwater Protection Project, 1987.
Resources for the Future, Inc. (RFF). National Pesticides Inventory Database. Washington,
D.C, 1985.
U.S. Department of Agriculture (USDA). The Magnitude and Costs of Groundwater
Contamination From Agricultural Chemicals. Washington, D.C.: Staff Report
AGES870318,1987.
U.S. Environmental Protection Agency, MADEQE, CCPEDC et al. "Cape Code Aquifer
Management Project" (CCAMP), Draft Document, 1988.
U.S. Environmental Protection Agency. Comparative Impact Analysis of Sources of
Ground-Water Contamination-Phase III, Draft Report Washington, D.C.: OPA,
January 1987.
U.S. Environmental Protection Agency. Unfinished Business: A Comparative Assessment of
Environmental Problems. Washington, D.C: OPA and OPPE, February 1987.
U.S. Environmental Protection Agency. Report to Congress: Class VInjection Wells;
Current Inventory. Washington, D.C: USEPA Office of Water, EPA 570/9-87-006, May
1987.
U.S. Environmental Protection Agency. Protecting Groundwater: Pesticides and
Agricultural Practices. Washington, D.C: EPA 440/6-88-001, OGWP (WH550G),
February 1988.
-77
-------
20. & 21. Pesticide Residues & Pesticide Application
Problem Area Definition
These two problem axeas address the health risk associated with pesticide residues on food
and pesticide application. Analysis of pesticide residues on food includes consideration of
residues on foods imported into the region as well as residues on food grown in the region.
Analysis of risks from pesticide application includes risk to professional applicators, home
applicators, and bystanders. This analysis also covers improper application and accidental
exposure.
Summary/Abstract
The risks from pesticide residues and applications are primarily calculated by apportioning
the national risk to Region I. For pesticide residues, the national risk (6,000 cancer cases) is -
apportioned based on die regional population, resulting in an estimated 317 annual cancer
cases in New England. The national cancer risk from pesticide application (100 cases) is
apportioned according to use of pesticides in the region. Using this method, less than one
annual cancer case is associated with application in Region I. Non-cancer effects are
associated only with the application of pesticides. Using hospital admission figures and a
pesticide poisoning incidence rate, approximately 27 applicators and farm workers are
admitted to hospitals for pesticide-related health effects. Impacts on home applicators and
bystanders are not assessed.
Toxicity Assessment
Specific pesticides are not addressed in this analysis. Risks are estimated based on the
national risk estimates calculated in Unfinished Business. For effects from pesticide
application, the acute risks are based on hospital admissions. These admissions are assumed
to be related to respiratory problems.
78
-------
Exposure Assessment
More than 200 pesticide chemicals are potentially oncogenic. Specific concentrations of
these pesticides are not estimated. The potential pathways of exposure include ingestion of
pesticide residues on food and direct inhalation of and contact with pesticides being applied.
The total regional population is assumed to be exposed to pesticide residues on food. No
calculation of total population exposed is made for pesticide application.
Risk Characterization
The cancer risk associated with pesticide residues is calculated by apportioning the
national risk (6,000 cases annually) from Unfinished Business to Region I, based on the
region's percentage of the U.S. population. This results in an annual incidence of 317 cases
for the total New England population. We did not estimate non-cancer risk associated with
pesticide residues on food.
The cancer risk from pesticide application is determined by apportioning the national
estimate (100 cases annually) to the region based on annual pesticide use. The result is less
than one annual cancer case.
For pesticide applications, the acute risks are calculated by multiplying the number of
hospital admissions in Region I (1.8 million) by a pesticide poisoning incidence rate. The
pesticide poisoning rate used from EPA's National Study of Hospital Admitted Pesticide
Poisonings is 0.6 per 100,000 people for applicators, and 0.9 per 100,000 people for farm
workers. This results in an annual incidence of 11 applicator and 16 farm worker
hospitalizations. The impacts on home applicators and bystanders are not assessed.
The uncertainty associated with the risks from pesticide residues and application is high.
The calculation of cancer cases from pesticide residues assumes that estimates from
Unfinished Business are accurate and that exposure across the region is equivalent. The
calculation of risk associated with pesticide application assumes that all pesticides are of equal
toxicity. The use of hospital data in this case may bias the results downward because it is
generally believed that a significant percentage of acute poisonings go unreported.
79
-------
22. Lead
Problem Area Definition
This category considers risks from exposure to lead in soil, drinking water, and air.
Summary/Abstract
The lead problem area ranks in the highest category for public health risk. The driving
health endpoint is learning disabilities in young children. The basis for this analysis is the
prevalence of confirmed cases of lead poisoning in young children as reported by the six New
England states. Recent data noted 1,951 confirmed cases of lead poisoning in New England,
and approximately 10,000 additional cases would be anticipated if all children less than six
years of age were screened for lead. Lead is not considered a carcinogen. The major
exposure routes for lead in New England are dust, paint and soil. Minor exposure routes
include drinking water and air.
Toxicity Assessment
Lead metal and its compounds are the materials of concern in this problem area. The
compounds include acidic lead (lead acetate), isocyanic acid (methyl isocyanate), lead
phosphate (phosphoric acid), lead stearate, and lead sulfide. For simplicity, the term "lead"
will be used to cover lead metal and the lead compounds.
Health Effects of Concern
Non-Cancer Health Effects
There is no convincing evidence that lead has any beneficial biological effect in humans.
(Expert Committee on Trace Metal Essentiality, 1983; and included in the Lead Criteria
document, 1986.)
Elevated blood-lead levels have been linked to a wide range of health effects, which are of
particular concern in young children. These effects range from relatively subtle changes in
biochemical measurements at 10 ug/dl (micrograms/deciliter) and below to severe retardation
and even death at very high levels (80 to 100 ug/dl). Lead can interfere with blood-forming
processes, vitamin D metabolism, kidney function, neurological processes, and reproductive
functions in both males and females. In addition, the negative impact of lead on cognitive
performance (as measured by IQ tests, performance in schools, and other means) is generally
80
-------
accepted at moderate-to-high blood-levels (30 to 40 ug/dl and above), and several studies also
provide evidence for possible attentional and IQ deficits at levels as low as 10 to IS ug/dl.
Changes in electroencephalogram readings, as another example, have also been observed at
low levels. For many subtle effects, the data may represent die limits of detectability of
biochemical and other changes and not necessarily actual thresholds for effects (Ronnie
Levin, Reducing Lead in Drinking Water: A Benefit Analysis, OPPE, USEPA, 1986).
This analysis uses learning disabilities as the non-cancer health endpoint of concern. This
endpoint is the health problem most frequently encountered in the most vulnerable
population, children less than six years of age. However, it is important to understand that the
evaluation of learning disabilities in young children will tend to understate the overall health
risks posed by lead. Severe lead intoxication can be fatal, for example. While this result is no
longer common Oead poisoning was a leading cause of death for American children in the
early part of this century), the risk remains. As indicated above, lead is associated with a
range of serious health problems in other populations, particularly women of cbildbearing age
and middle-aged males. Unfortunately data are insufficient on lead exposure or incidence of
lead poisoning for these groups. Since this analysis does not consider the rare cases of severe
lead intoxication or the effects of lead on all segments of society, it represents a lower bound
on the public health risk posed by lead in New England.
Cancer Health Effects
Lead is classified as a Group B 2 carcinogen by EPA, indicating that there is some
evidence of carcinogenicity from animal studies. However, several epidemiology studies
have not shown an association between lead exposure and cancer incidence. This analysis
does not consider cancer risks posed by lead.
Exposure Assessment
The data used in this analysis are prevalence data on lead poisoning in children in New
England.
The Center for Disease Control (CDC) defines lead poisoning as 25 micrograms per
deciliter (ug/dl) of lead in blood with free erythrocyte protoporphyrin (FEP) greater than
35 ug/dl. FEP is used as an indicator of potential lead poisoning.
Most New England communities conduct some lead screening for children less than six
years of age. Screening programs take a capillary blood sample from a child and analyze the
sample for FEP. If the FEP is greater than 35 ug/dl, the parents are contacted and a venous
blood sample is taken from the child. This sample is analyzed for lead. If this blood lead is
greater than 25 ug/dl, the child is counted as a confirmed case of lead poisoning.
81
-------
State health departments maintain the results of screening programs but there is no .
common format for recording or reporting this information. All six New England state health
departments were contacted for this analysis. Massachusetts, Maine, New Hampshire, and
Rhode Island reported the percentage of children screened and the number of confirmed cases
of lead poisoning. Connecticut reported the percentage screened and the percentage of
children presumed at risk. Vermont reported that both targeted and broad-based screening
efforts in recent years have identified only one confirmed case of lead poisoning.
The following table summarizes the information received from state health departments.
State Number of Children Screened Cases
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
-
4,151
165,852
7,150
17,455
Screening
(£12%)
(4%)
(39%)
(10%)
(24%)
data not reported
(6,060 potential)
249
1,235
39
427
1
Pathways of Exposure
This analysis does not isolate any particular exposure route or set of exposure routes for
lead. Rather it relies on data on the prevalence of lead poisoning, which reflects exposure
from all pathways.
Lead is a ubiquitous environmental contaminant It is found in air, water, and soil as well
as in plumbing, housepaint, food, and dust This section discusses the exposure routes of lead
and the risks they present
As stated above, lead is found in every medium throughout the environment Virtually all
lead contamination is the result of human activity. Furthermore, lead moves easily from one
medium to another. The major contributors to lead poisoning in New England are lead paint,
lead soil, and lead dust Minor but still significant exposure routes are lead in air and lead in
drinking water.
Lead paint is common in homes constructed before 1950. As many as 85 percent of these
homes may contain lead paint The age and condition of the housing stock are predictors of
lead poisoning. Lead poisoning cases are most often found in older inner city neighborhoods.
However, cases are also routinely reported in suburban and rural communities. Children are
at risk wherever they are exposed to lead.
82
-------
Cases of acute lead poisoning (greater than 40 ug/dl) are usually the result of ingestion of
lead paint These cases are less common than in the past.
The most common cases are that of blood lead between 25 ug/dl and 40 ug/dl. Recent
research indicates that inhalation or ingestion of lead-contaminated dust is the critical
exposure pathway for these cases. The relationships between paint, soil, and dust are not well
understood. Researchers believe that soil is contaminated with lead from deteriorating interior
and exterior lead paint, from deposition of airborne lead, and from lead in water. All of these
sources contribute to the contamination of dust. It is important to note that lead does not
leach through soil. Rather, it remains where it was deposited. As a result, surface soil
contains the accumulation of years of lead pollution from gasoline, old paint, etc. The soil
acts as a reservoir of lead contamination.
CDC's most recent statement on lead poisoning (1985) asserts that soil contaminated at
500 to 1,000 parts per million (ppm) lead or more is associated with significant increased risk
for lead poisoning in children. Data on soil contamination in residential areas is scarce.
Sampling conducted in Boston in October 1987 identified average surface soil levels greater
than 1,900 ppm lead.
The reduction of lead in gasoline has led to a dramatic reduction in airborne lead.
However, most of the airborne lead from past years is probably now deposited as lead in soil.
Lead in drinking water remains a problem. Recent tests in Boston revealed that more than
half of the city schools have unsafe levels of lead in their drinking water, primarily as a result
of slightly acidic water interacting with old lead pipes and lead solder in plumbing fixtures.
The lead levels in drinking water are unlikely to cause lead poisoning in the absence of other
sources, but they do contribute to the overall body burden of lead. This in turn increases the
risk of lead poisoning from other sources.
Lead in drinking water may be a major source of concern for adults. Adults tend to drink
more water than children and because of behavior patterns are likely to be at lower risk from
exposure to soil, dust, and paint
Population at Risk
The population at risk is all children under six years of age.
State Children Under Six Years
Connecticut 202,000
Maine 103,770
Massachusetts 430,143
New Hampshire 71,084
Rhode Island 72,842
Vermont 47,459
Total 927,298
83
-------
Risk Characterization
An estimate of lead poisoning cases in New England was determined by extrapolating
from existing screening and prevalence data to project the number of cases if all children
under six years were screened.
The first step in the calculation was to total all of the reported confirmed cases of lead
poisoning, giving a total of 1,951.
The next step was to extrapolate from the known cases to the number of expected cases if
screening programs covered 100 percent of the children under six. The analysis used a
weighted average to determine expected cases in Maine and Rhode Island because the high
rates in these states seemed to reflect the screening of the most vulnerable children. Vermont
was not used in the calculation of the weighted average. The rate (#cases/#screened) for each
state was multiplied by the number of children screened in each state as a percentage of the
total number of children screened in New England. The resulting weighted fractions for each
state were then summed to get a weighted rate of 1.2 percent The weighted rate was used to
determine expected cases in Maine, Rhode Island, and Vermont. For Massachusetts and New
Hampshire, expected cases were determined by using the rate for that state. For Connecticut,
the 3 percent rate was used.
The final step was to combine reported cases with expected cases.
Connecticut: 6,060 (3% of children under six)
Maine: 1,444 (249 reported +1,195 expected)
Massachusetts: 3,191 (1,235 reported + 1,956 expected)
New Hampshire: 391 (39 reported + 352 expected)
Rhode Island: 1,092 (427 reported + 665 expected)
Vermont: 571 (1 reported + 570 expected)
Total: 12,749 (1,951 reported + 10,798 expected)
Projected Overall New England Rate: 1.37 percent
These numbers are not meant to reflect the precise number of cases that exist in New
England or in any individual state. Rather, they reflect the approximate magnitude of health
risk to children from lead poisoning.
When this calculation is placed in the format of the Comparative Risk Project, it produces
the following results.
Population Score: 3
Health Endpoint Score: 3
Ratio Score: 4
Total Score 10
84
-------
Using a different methodology for calculating the expected cases of lead poisoning would
not change the scoring results. The only factor that is not fixed is the ratio score. Using only
the 1,951 confirmed cases in New England, the rate is 10-2 (1,951 cases/194,608 children
screened). This value still results in a ratio score of 4.
One test of the validity of these results is comparison with other studies. The data from the
Second National Health and Nutrition Examination Survey (NHANES n) (1976-1980)
indicate that 4 percent of American children less than five years old have blood-lead levels at
or above 30 ug/dl. This implies that the rate developed in this analysis, 1.37 percent; is likely
to be a lower bound. The actual rate, therefore, may be higher.
The uncertainty in this analysis is low. The data collected are reliable and the calculations
are simple. The estimated number of lead poisoning cases-12,749—is believed to be the
correct order magnitude.
Percentage of Problem Covered
The percentage of problem covered for lead exposure in young children is high, and this is
widely accepted as the most vulnerable population. As indicated above, there are other
populations at risk from lead, including women of childbearing age and middle-aged men.
However, since exposure data for these groups are limited and lead is ranked in the highest
residual public health risk category without including an analysis of these groups, further
analysis was not conducted
85
-------
23. Asbestos
Problem Area Definition
Asbestos has become an issue of great environmental concern because of the
epidemiological evidence showing a strong correlation between cumulative asbestos
exposures and deaths. Asbestos is associated with such lethal diseases as lung cancer and
mesothelioma (a relatively rare cancer of the thin membranes lining the chest and the
abdomen). Asbestos has also been linked to gastrointestinal (GI) cancers and cancers at other
anatomical sites, including the larynx, kidney, and ovary.
Asbestos can be released to the ambient air at numerous stages in its production and use.
There are three general types of asbestos exposure:
• Occupational (e.g., worker exposures at facilities involved in mining and milling,
fiber processing into products, installation of products, building maintenance)
• Non-occupational (e.g., exposures resulting from time spent in
asbestos-containing buildings, such as schools, office buildings, and residences)
• Ambient (e.g., general background levels resulting from construction and
demolition of buildings containing asbestos, asbestos removal; in some areas,
elevated background levels may be found because of the existence of either large
manufacturing operations using asbestos or asbestos mining and milling
facilities)
This evaluation developed rough approximations of the risks to human health resulting
from the latter two exposure scenarios.
Summary/Abstract
Based on the results of a risk assessment completed by EPA's Office of Health and
Environmental Assessment, we developed screening-quality estimates of individual risk and
annual incidence for three cancer types (lung cancer, mesothelioma, and GI cancer) and three
exposure scenarios (rural ambient air--0.07 ng/m3, urban ambient air-3 ng/m3, and
non-occupational/indoor air--80 ng/m3). The risk assessment considered only exposures
through inhalation. Because of data limitations on the number of individuals exposed to
asbestos in Region I, we assumed conservatively that the total Region I population was
exposed to asbestos in the ambient air-subdivided by residence in either rural or urban
areas-and in non-occupational settings. This approach allowed us to develop a range of
incidence projections that we feel bounds the problem for screening purposes.
86
-------
The estimated lifetime individual cancer risks for non-occupational exposures ranged from
1E-5 (GI cancer) to 1E-4 (mesothelioma and lung cancer). The estimated lifetime individual
cancer risks for urban and rural ambient exposures were, respectively, one and two orders of
magnitude smaller. The analysis predicted approximately five annual cases of cancer from
ambient air exposures. The urban exposures accounted for almost all of the estimated
incidence (approximately 99 percent). The estimated annual cancer incidence associated with
non-occupational exposures was significantly higher, contributing roughly 178 annual cancer
cases.
Toxicity Assessment
Asbestos has a strong association with lung cancer and mesothelioma. It has also been
linked to GI cancers and cancers at the larynx, kidney, and ovary; however, the degree of risk
and the strength of the evidence supporting these diseases are less than for lung cancer or
mesothelioma. The International Agency for Research on Cancer lists asbestos as a Group 1
carcinogen; EPA's guidelines characterize asbestos as a Group A human carcinogen.
To quantify the risks associated with lung cancer and mesothelioma, we relied on the risk
calculations presented in the EPA document Airborne Asbestos Health Assessment Update,
EPA/600/8-84/003F, June 1986. The available data suggest that the excess risk of lung
cancer from asbestos is proportional to the cumulative exposure (the duration times the
intensity) and the underlying risk in the absence of asbestos exposure. The risk of death from
mesothelioma is approximately proportional to the cumulative exposure to asbestos and
increases exponentially with time since the initial exposure. This analysis assumed
conservatively a lifetime exposure starting at birth; it did not consider smoking habits, which
can increase the risks. Thus, the analysis used the following approximate cancer unit risk
factors:
• 9.3E-6 (f/ml)-l for mesothelioma (women)
• 6.3E-6 (f/ml)-l for mesothelioma (men)
' • 1.7E-6 (f/ml)-l for lung cancer (women)
• 5.7E-6 (f/ml)-l for lung cancer (men)
87
-------
The information available is insufficient to quantify the dose-response relationships for
other asbestos cancers. However, excess GI cancer mortality is approximately 10 to 30
percent of the risk attributable to lung cancer. We calculated GI cancer risks using a midpoint
of 20 percent
Asbestos is also associated with non-cancer effects. Exposure to asbestos may increase the
risk of asbestosis, a chronic lung ailment resulting in shortness of breath, permanent lung
damage, and an increased risk of lung infections. However, there was insufficient information
to estimate the occurrence of asbestosis in this preliminary risk screen.
Exposure Assessment
Asbestos can be released at numerous points in its production and use. This analysis
considered only ambient air and non-occupational exposures. Because these points of
exposure are ubiquitous, we relied on available monitoring data to characterize average
concentrations found in each exposure setting. Table 23-1 presents the mean measured
concentration values from studies (both ambient and indoor) deemed by EPA to have a
reasonable number of sampling episodes.
Ambient air exposures were divMed into two categories: rural and urban. An average
ambient air concentration of 3 ng/m3 was used to predict cancer risks in urban environments.
The city monitoring data in Table 23-1 generally support this value, with the exception of the
findings for Paris, France. Also, EPA's 1985 analysis of hazardous air pollutants used an
average asbestos ambient level of 3 ng/m3 in its analysis of cancer risks.1
Because large fractions of the population in three Region I states (Vermont, New
Hampshire, and Maine) reside in rural areas, it was important to consider asbestos levels
found in more remote areas. Atmospheric sampling programs conducted in rural areasjn the
United States and Germany detected asbestos fiber levels between 0.01 and 0.12 ng/m3
(corresponding to roughly 3E-7 f/ml and 4E-6 f/ml, respectively). A midpoint of 0.07
ng/nr was used as the representative concentration for rural areas.
*U.S. EPA, Office of Policy Analysis, An Analysis of Carcinogenic Risks for Selected Pollutants Nationwide,
May 1985.
^Federal Register 40 CFR Part 763. "Asbestos; Proposed Mining and Import Restrictions and Proposed
Manufacturing Importation and Processing Prohibitions," January 29,1986.
-------
Table 23-1
Summary of Environmental Asbestos Sampling
Sample Set
Collection
Period
Number of
Samples
Mean
Concentration
ng/m
Quarterly composites of 5 to 7
24-hour U.S. samples
(Nicholson, 1971; Nicholson
and Pundsack, 1973)
Quarterly composite of 5 to 7
24-hour U.S. samples (U.S.
EPA, 1974)
5-day samples of Paris, France
(Sebastian etal., 1980)
6- to 8-hour samples of New York
City (Nicholson at al., 1971)
5-day 7-hour control samples for
U.S. school study (Constant et
al., 1983)
16-hour samples of 5 U.S. cities
(U.S. EPA, 1974)
New Jersey schools with damaged
1969-1970
1969-1970
187
127
3.3Ca
3.4C
pupil-use areas (Nicholson et
al., 1978)
U.S. school rooms/areas with 1980-1981
asbestos surfacing materials
(Constant, 1983)
U.S. school rooms/areas with 1980-1981
flchoctne cnrf9/*inn mfltorislc
ttauoMUo oUMavliiy iilcuoiicuo
(Constant, 1983)
Buildings with asbestos materials in 1976-1977
Paris. France (Sebastien et al.,
1980)
U.S. buildings with friable asbestos 1974
in plenums or as surfacing
materials (Nicholson et al.,
1975; Nicholson et al., 1976)
U.S. buildings with cementitious 1974
asoesios maienai in plenums or
as surfacing materials
(Nicholson et al.. 1975, 1976)
Ontario buildings with asbestos 1982
insulation (Ontario Royal
Commission, 1984)
54 183(179C,4A)
31 61(53C,8A)
135 35(25C,10A)
54 48C
28 15C
63 2.1
*C m Chrysolite.
DA _ AmnKIKMlA
89
-------
An average asbestos concentration level of 80 ng/m3 was used to characterize
non-occupational exposures. This value was derived by averaging the monitoring results
presented in Table 23-1 for schools and public buildings.
As a cautionary note, it is important to mention that the measured asbestos levels
summarized in Table 23-1 were taken during a period of time that may not adequately reflect
current conditions. Given the recent regulatory activity and the increased public awareness of
the risks posed by asbestos, current levels could be lower, thus resulting in lower risk
estimates. The risk screen may also result in overestimates of risk because data limitations
made it necessary to assume conservatively that all asbestos exposures started at birth and
continued for a lifetime. The individual risk results presented can be easily adjusted to reflect
alternate concentration levels should more recent data become available.
The primary exposure pathway for asbestos is inhalation. Some studies have also
suggested the possibility of adverse human health effects through ingestion of asbestos fibers,
but this route of exposure is still under investigation. To date, there is no evidence of disease
incidence resulting from dermal exposures.
Because of data limitations on the population exposed to asbestos, we assumed
conservatively that the total Region I population was exposed to the range of asbestos
concentration levels (ambient and non-occupational) over the course of a lifetime, starting at
birth. This approach allowed us to develop a range of incidence projections that we feel
bounds the problem for screening purposes. U.S. census data were used to identify the total
population of men and women residing in Region I. Census data were also used to determine
the proportion of men and women residing in metropolitan and nonmetropolitan areas for
each Region I state to complete the excess cancer incidence calculations for ambient air
exposures. The population data used in this analysis are summarized in Attachment 23-B.
Risk Characterization
The lifetime individual cancer risks for non-occupational exposures could range as high as
1E-4, whereas ambient exposures are two to four orders of magnitude lower. Ambient urban
exposures are associated with total lifetime individual cancer risks of approximately 1E-S; the
total lifetime individual cancer risk for remote rural areas is approximately 1E-7. Table 23-2
summarizes the estimated lifetime individual cancer risk estimates for each concentration
scenario and health effect category.
90
-------
Table 23-2
Estimated Lifetime Individual Cancer Risks for Asbestos
Health Effect
Ambient Ambient Non-Occupational
Rural Exposures1 Urban Exposures2 Exposures
(0.07 ng/m3) (3 ng/m3) (80 ng/m3)
Mesothelioma:
- Males
- Females
Lung Cancer:
- Males
- Females
Gl Cancers:
- Males
- Females
4.4E-07
6.5E-07
4.0E-07
1.2E-07
7.9E-08
2.3E-08
1.9E-05
2.8E-05
1.7E-05
5.0E-06
3.4E-06
1.0E-06
5.1E-04
7.5E-04
4.5E-04
1.4E-04
9.1E-05
2.9E-05
Note: These screening-quality risk results are for use only in establishing a relative ranking of
environmental issues for further study.
1 Based on information presented in 40 CFR Part 763, January 29,1986. remote rural
concentrations were found to range between 0.01 ng/m3 and 0.12 ng/m3. For this analysis,
a midpoint of 0.07 ng/m3 was used. Risk results were linearly extrapolated from the EPA risk
results.
2 Source: U.S. EPA, Office of Health and Environmental Assessment, Airborne Asbestos
Health Assessment Update. EPA/600/8-84/003f, June 1986, p. 165.
3 Estimated as 20 percent of the predicted lung cancer risks.
91
-------
Table 23-3 presents the predicted annual incidence of cancer associated with ambient and
non-occupational exposures. As shown, total ambient exposures (rural and urban) contribute
approximately five excess cancer cases annually. The incidence estimates for
non-occupational exposures are more than two orders of magnitude higher, potentially
reaching as high as 176 annual excess cancer cases.
To properly interpret and use the asbestos risk screen results in the risk evaluation project,
it is important to understand the limitations of the risk screen methodology. These
uncertainties are grouped based on whether they are expected to yield overestimates or
underestimates of risk
• Assumptions/data limitations resulting in overestimates of risk
- Lifetime asbestos exposures start at birth. If better site-specific information
existed on this topic, we could revise the risk estimates to reflect more
realistic exposure levels.
- All individuals in Region I are exposed to asbestos. This is a worst-case
assumption; better site-specific information on populations exposed at each
concentration level would probably result in lower risk estimates.
- Risk projections are based on concentration data from relatively old studies.
Collection of new information would provide better indications of current
asbestos concentration levels given recent regulatory activity and increased
public awareness. Also, use of site-specific monitoring data would help.
- Extrapolation from high-dose to low-dose exposures is necessary. There are
significant uncertainties in developing risk estimates for low concentration
levels based primarily on high-level occupational exposure studies. This is a
concern for all risk assessments.
Assumptions/data limitations resulting in underestimates of risk
- Only exposure through inhalation was considered. There is some concern
that ingestion of asbestos could pose risks. Studies are currently being
conducted to explore this topic further.
- Lung cancer risks for smokers are significantly higher. EPA has generated
individual cancer risk estimates with and without smoking habits. Because of
limited information on smokers and their exposure levels in Region I, we
based the risk screen on the individual cancer risks that did not consider
smoking habits. These risk estimates fell between the individual risks
calculated by EPA for smokers and nonsmokers (see Attachment 23-A).
92
-------
- Exposure levels in some buildings could be higher than 300 ng/nr. For
example, while the mean of the measured asbestos levels in the New Jersey
schools sampled (Nicholson et al., 1978, see Table 23-1) was.217 ng/m3, the
observed asbestos levels ranged from 9 ng/m3 to 1,950 ng/m3.
We believe that the risk assessment methodology characterized in this paper has captured
most of the problem. However, the uncertainty in the risk estimates for the ambient exposures
is rated medium and the uncertainty for the non-occupational exposures is rated high.
Table 23-3
Estimated Annual Cancer Incidence for Asbestos
Health Effect
Ambient
Rural Exposures''
(0.07 ng/m3)
Ambient Non-occupational
Urban Exposures' Exposures
(3 ng/m3) (80 ng/m3)
Mesothelloma:
- Males
- Females
Lung Cancan
- Males
- Females
Gl Cancers:
- Males
- Females
0.01
0.01
0.01
0.00
0.00
0.00
1.34
2.12
1.20
0.38
0.24
0.08
44.19
69.92
39.54
13.23
7.91
2.75
Total
0.03
5.35
177.53
Note: These screening-quality risk results are for use only In establishing a relative ranking of
environmental Issues for further study.
1 Based on information presented in 40 CFR Part 763, January 29,1986, remote rural concentrations
were found to range between 0.01 ng/m3 and 0.12 ng/m3. For this analysis, a midpoint of 0.07 ng/m3
was used. Risk results were linearly extrapolated from the EPA risk results.
2 Source: U.S. EPA, Office of Health and Environmental Assessment, Airborne Asbestos Health
Assessment Update, EPA/600/8-84/003f, June 1986, p. 165.
3 Estimated as 20 percent of the predicted lung cancer risks.
93
-------
Attachment 23-A
-------
Table 23-A-1
Lifetime Risks per 100,000 Persons of Death from Mesothelloma and
Lung Cancer from Continuous Asbestos Exposures of 0.0001 and 0.01 I/ml
According to Age and Duration of Exposure. U.S. General Population Death
Rates Were Used and Smoking Habits Were Not Considered. '
Age at
Onset of
Exposure
Concentration • 0.0001 I/ml
Years of Exposure
10
20
Life-
time
Concentration = 0.01 I/ml
Years of Exposure
10
20
Life-
time
Mesothelloma In Females
0
10
20
30
50
0.1
0.1
0.1
0.0
0.0
0.7
0.4
0.3
0.1
0.0
1.2
0.8
0.4
0.2
0.0
2.0
1.2
0.7
0.3
0.0
2.8 14.6
1.5 9.4
0.8 5.6
0.4 3.1
0.0 0.6
67.1 120.8 196.0 275.2
42.6 75.5 118.7 152.5
25.1 43.5 65.7 78.7
13.3 22.4 31.9 35.7
2.1 3.2 3.9 3.9
Lung Cancer In Females
0
10
20
30
50
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.1
0.2
0.2
0.2
0.2
0.1
0.5 1.0
0.4 1.0
0.3 1.0
0.3 1.0
0.1 0.7
4.6
4.6
4.6
4.6
3.1
9.2
9.2
9.2
9.0
5.5
18.5
18.6
18.2
16.7
8.1
52.5
43.4
34.3
25.1
8.8
Mesothelloma In Males
0
10
20
30
50
0.1
0.1
0.0
0.0
0.0
0.5
0.3
0.2
0.1
0.0
0.9
0.6
0.3
0.1
0.0
1.5
0.8
0.4
0.2
0.0
1.9 11.2
1.1 7.0
0.5 4.1
0.2 2.1
0.0 0.3
61.0 91.1 145.7 192.8
31.2 58.2 84.7 106.8
17.5 30.1 44.5 51.7
8.8 14.6 20.4 22.3
1.1 1.8 2.0 2.1
Lung Cancer In Males
0
10
20
30
50
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.2
0.1
0.1
0.3
0.3
0.3
0.3
0.2
0.6
0.6
0.6
0.6
0.3
1.7
1.4
1.1
0.8
. 0.3
2.9
2.9
3.1
3.1
2.5
14.8
14.9
15.0
14.9
11.5
29.7
29.8
30.0
29.8
20.3
59.2 170.5
59.5 142.0
59.4 113.0
56.6 84.8
29.1 30.2
1The 95 percent confidence limit on the risk values for lung cancer for an Unstudied exposure circumstance is a factor of 10.
The 95 percent confidence limit on the risk values for lung cancer on the average determined from 11 unit exposure risk
studies is a factor of 2.5. The 95 percent confidence limit on the risk values for mesothelioma for an unstudied exposure
circumstance is a factor of 20. The 95 percent confidence limit on the risk values for mesothelioma for a studied
circumstance can be reasonably averaged as a factor of 5. The values for continuous exposure were derived by
multiplying 40 hrtwk risks, obtained from occupational exposures, by 4.2 (the ratio of hours in a week to 40 hours).
-------
Table 23-A-2
Lifetime Risks per 100,000 Females of Death from Mesothelioma and
Lung Cancer from Continuous Asbestos Exposures of 0.0001 and 0.01 I/ml
According to Age at First Exposure, Duration of Exposure, and Smoking'
Concentration = 0.0001 I/ml Concentration = 0.01 Vm\
Years of Exposure Years of Exposure
Age at
Onset of Life- Life-
Exposure 1 5 10 20 time 1 5 10 20 time
Mesothelloma In Female Smokers
0 0.1 0.6 1.2 1.9 2.5 13.9 64.0 115.1 186.2 252.0
10 0.1 0.4 0.7 1.1 1.4 9.0 40.3 71.4 112.0 142.8
20 0.1 0.2 0.4 0.6 0.7 5.3 23.5 40.7 61.3 72.8
30 0.0 0.1 0.2 0.3 0.3 2.8 12.3 20.6 29.4 32.8
50 0.0 0.0 0.0 0.0 0.0 0.6 2.0 2.9 3.5 3.5
. Lung Cancer In Female Smokers
0 0.0 0.1 0.3 0.5 1.5 2.8 13.4 26.7 53.3 149.9
10 0.0 0.1 0.3 0.5 1.2 2.8 13.4 26.7 53.3 123.5
20 0.0 0.1 0.3 0.5 1.0 2.8 13.4 26.7 52.5 96.9
30 0.0 0.1 0.3 0.5 0.7 2.8 13.3 25.9 47.9 71.0
50 0.0 0.1 0.2 0.2 0.2 2.0 6.8 15.5 22.7 24.4
Hesothelloma In Female Nonsmokers
0 0.1 0.7 1.2 2.0 2.7 14.8 68.2 122.8 199.4 272.2
10 0.1 0.4 0.8 1.2 1.6 9.5 43.4 81.2 121.2 155.8
20 0.1 0.3 0.4 0.7 0.8 5.7 25.6 44.4 67.2 80.6
30 0.0 0.1 0.2 0.3 0.4 3.1 13.6 23.0 32.9 36.8
50 0.0 0.0 0.0 0.0 0.0 0.6 2.2 3.4 4.1 4.1
Lung Cancer In Female Nonsmokers
0 0.0 0.0 0.0 0.1 0.2 0.3 1.3 2.7 5.2 16.4
10 0.0 0.0 0.0 0.1 0.1 0.3 1.3 2.7 5.3 13.9
20 0.0 0.0 0.0 0.1 0.1 0.3 1.3 2.7 5.2 11.3
30 0.0 0.0 0.0 0.1 0.1 0.3 1.3 2.7 5.0 8.7
50 0.0 0.0 0.0 0.0 0.0 0.3 1.1 2.1 3.5 3.9
1The 95 percent confidence limit on the risk values for lung cancer for an unstudied exposure circumstance is a factor of 10.
The 95 percent confidence limit on the risk values for lung cancer on the average determined from 11 unit exposure risk
studies is a factor of 2.5. The 95 percent confidence limit on the risk values for mesothelioma tor an unstudied exposure
circumstance is a factor of 20. The 95 percent confidence limit on the risk values for mesothelioma for a studied
circumstance can be reasonably averaged as a factor of 5. The values for continuous exposure were derived by
multiplying 40 hr/wk risks, obtained from occupational exposures, by 4.2 (the ratio of hours in a week to 40 hours).
-------
Table 23-A-3
Lifetime Risks per 100,000 Males of Death from Mesothelioma and
Lung Cancer from Continuous Asbestos Exposures of 0.0001 and 0.01 I/ml
According to Age at First Exposure, Duration of Exposure, and Smoking'
Age at
Onset of
Exposure
Concentration = 0.0001 fAnl
Years of Exposure
10
20
Life-
time
Concentration = 0.01 f/ml
Years of Exposure
10
20
Life-
time
Meaothelloma In Male Smoker*
0
10
20
30
50
0.1
0.1
0.0
0.0
0.0
0.5
0.3
0.2
0.1
0.0
0.9
0.5
0.3
0.1
0.0
1.4
0.8
0.4
0.1
0.0
1.8
1.0
0.5
0.2
0.0
10.6
6.6
3.6
2.0
0.3
48.3 85.5 137.5 181.0
29.4 51.5 77.8 98.3
16.4 28.0 41.2 47.9
8.1 13.4 18.5 20.2
1.1 1.5 1.8 1.8
Lung Cancer In Male Smokers
0
10
20
30
50
0.0
0.0
0.0
0.0
0.0
0.2
0.2
0.2
0.2
0.2
0.4
0.4
0.4
0.4
0.3
0.8
0.8
0.8
0.8
0.4
2.4
2.0
1.6
1.2
0.4
4.2 20.9
4.2 21.0
4.2 21.3
4.2 21.3
3.6 16.2
41.9
42.0
42.3
42.0
28.4
83.4 238.1
83.9 197.8
83.4 157.5
79.2 117.6
40.3 42.0
Mesothelioma In Male Nonsmokero
0
10
20
30
50
0.1
0.1
0.0
0.0
0.0
0.6
0.4
0.2
0.1
0.0
1.0
0.6
0.4
0.2
0.0
1.6
1.0
0.5
0.2
0.0
2.2
1.2
0.6
0.3
0.0
12.5 57.0
7.8 35.3
4.5 20.4
2.4 10.5
0.4 1.5
2.3 164.5 220.1
2.6 97.3 122.6
5.1 52.4 61.7
7.5 24.6 26.9
2.2 2.7 2.7
Lung Cancer In Male Nonsmokera
0
10
20
30
50
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.0
0.2
0.2
0.1
0.1
0.0
0.3
0.3
0.3
0.3
0.3
1.5
1.5
1.5
1.5
1.3
2.9
2.9
2.9
2.9
2.2
5.9
5.9
5.9
5.7
3.9
18.5
15.5
12.5
9.7
4.2
1The 95 percent confidence limit on the risk values for lung cancer for an unstudied exposure circumstance is a factor of 10.
The 95 percent confidence limit on the risk values for lung cancer on the average determined from 11 unit exposure risk
studies is a factor of 2.5. The 95 percent confidence limit on the risk values for mesothelioma for an unstudied exposure
circumstance is a factor of 20. The 95 percent confidence limit on the risk values for mesothelioma for a studied
circumstance can be reasonably averaged as a factor of 5. The values for continuous exposure were derived by
multiplying 40 hr/wk risks, obtained from occupational exposures, by 4.2 (the ratio of hours in a week to 40 hours).
-------
Attachment 23-B
-------
Table 23-B-1
Region I Population Data1
(1985 data, in thousands)
Region I
State Metropolitan Nonmatropolitan Total
Maine
New Hampshire
Vermont
Massachusetts
Rhode Island >,
Connecticut
420
560
122
5.291
896
2.940
744
438
4-13
531
72
233
1.164
998
535
5.822
968
3.173
Total 10.229 ~T4§T 12.660
1 Source: U.S. Department of Commerce, Bureau of the Census. Statistical
Abstract of the United States. 1987.
-------
Table 23-B-2
Percentage of Males and Females
In Each Region I State'
(1980 data)
Region I Percentage Percentage
Slate of Males of Females
Maine 49% 51%
New Hampshire 49% 51%
Vermont 49% 51%
Massachusetts 48% 52%
Rhode Island 48% 52%
Connecticut 48% 52%
1 Source: 1987 Statistical Abstract of the United States. A breakdown
of population by sex was available only for 1980; this breakdown
was assumed to remain unchanged for 1985.
-------
Table 23-B-3
Region I Population Data by Sex and Area1
(1985 data, in thousands)
Metropolitan Nonmetropolltan
Region I State
Maine
New Hampshire
Vermont
Massachusetts
Rhode Island
Connecticut
Total
Mate
206
274
60
2.540
430
1.411
4.921
Female
214
286
62
2.751
466
1.529
5.308
Male
365
215
202
255
35
112
1.184
Female
379
223
211
276
37
121
1.247
Total
1.164
998
535
5,822
968
3.173
12.660
1 Source: U.S. Department of Commerce, Bureau of the Census. Statistical Abstract of the
UnitedStatas, 1987.
------- |